Skip to main content

Full text of "History of Civilizations of Central Asia"

See other formats




S^B - 1 

VthliiniL- III 
I In crossroads of civilisations: a.d. 250 u> 


. " • 













n . 


UNESCO V ■ ,1 li I 1 s h 

i n a 

ISBN 978-92-3-103211-0 Contents CONTENTS 





alphabetical order) 17 


1993) 18 



Other collaborating specialists 22 


The Sasanians 24 

The Guptas 26 

The Sui dynasty 27 

The T'ang dynasty 28 

Oasis states 29 

Ecology, geography and climate 29 

Nomadic societies 31 

The Hsiung-nu and the Huns 33 

The Silk Route 35 

Cross-cultural influences 37 





ISBN 978-92-3-103211-0 Contents CONTENTS 

Struggles against the northern nomads 43 

The Sasanian administration 45 

Royal cities 45 

The reforms of Khusrau I 47 

The economy 48 

Sasanian coins 49 

The army, warfare and armaments 57 


Sasanian cities and fortifications 60 

Sasanian court architecture 63 

Sasanian religious architecture 66 

Sasanian art and crafts 67 


Written works 82 

Secular literature 82 

Royal Res Gestae 86 

Religious literature 88 

Philosophy and theology 90 

Visionary and apocalyptic texts 90 

Andarz (wisdom literature) 91 

Languages and scripts 91 


Borrowings and influences 94 

Science and philosophy 96 

Court chronicles and epic histories 99 

Religious life 102 

Religion and the law 105 


State organization and administration 107 

Economy, society and trade 112 

Religious life 113 

Cities, architecture, art and crafts 115 

Languages and scripts 121 




ISBN 978-92-3-103211-0 Contents CONTENTS 

Origin and rise of the Kidarites 123 

Early history of the Kidarites 124 

Foundation of the Kidarite state 125 

Kidara's conquest of Gandhara and Kashmir 126 

The decline of the Kidarites 127 

The new conquest of Bactria by the Sasanians (467) 129 

Economy, society and polity 130 

Monetary system and trade 132 

Life-style, culture and ideology 136 


Origins 138 

Political and military history 141 

Conquests in Gandhara and northern India 144 

The Hephthalites of Central Asia 146 

Social structure and administration 147 

Religious life and polyandry 150 

Language and scripts 151 

Towns 152 

Architecture 153 

Art and crafts 154 


MIR 167 

Eastern Kushans 167 

Kidarites in Gandhara and Kashmir 171 

Later Hephthalites 173 

Economic and cultural progress 176 

Sanskrit references 178 

Coinage of the Hunas 179 


The Khalaj, the successors of the Hephthalites 184 

Urban life and art in Tokharistan 186 



ISBN 978-92-3-103211-0 Contents CONTENTS 

Origin and political history of the Guptas 188 

Social and economic conditions 193 

Administration 197 

Religious life 199 

Literature 201 

Science 202 

Art and architecture 203 



Uncertain early history 212 

Social structure and administration 215 

Art, architecture, religion and language 222 


The Khwarizmian calendar 227 

Khwarizmian eras 229 

Religious beliefs 230 

The Arab conquest 233 

10 SOGDIANA 237 


Economic, cultural and social life 242 

Art and architecture 244 

Religious life 249 

Scripts, epics and literary sources 255 


Ustrushana 262 

Ferghana 275 

Chachandllak 278 


Geography and climate 282 

Peoples and languages 284 

Social life and the economy 286 

Shan-shan: administrative system 288 

Political upheavals 289 

Trade 291 


ISBN 978-92-3-103211-0 Contents CONTENTS 

Religion 292 

12 KOCHO ( KAO-CH'ANG) 298 

The Kocho Prefecture period (327-460) 300 

The Kocho kingdom period (460-640) 301 

Recently discovered documents at Kocho 302 

Administrative System and socio-economic life under the House of Ch'ii . . . 303 

The meeting and merging of cultures 305 


K'ang-chu 311 

The Huns 312 

The Hsien-pi 313 

Nomad kingdoms of northern China 315 

The Juan-juan 316 

The Turks 318 

The T'ieh-le and Kao-chii 318 

Archaeological evidence 320 


THE FIRST TURK EMPIRE (553-682) 322 

Ethnogenesis 322 

The economy 326 

Political history 327 


Resurgence of the Turk Empire 330 

Political and social structure 331 

Relations with China 333 

The empire in crisis 334 

The last war with T'ang China 335 

The final decade 336 

Epigraphic memorials of the Turks 337 

The Tiirgesh state 341 

The Uighurs and the Karluks 342 



ISBN 978-92-3-103211-0 Contents CONTENTS 

Three great empires 343 

The Western Regions under the Early T'ang 344 

The kingdom of Tibet 351 


(650-750) 358 


Tokharistan 362 

Kapisa- Gandhara 365 

Zabulistan 371 

The fight for independence 372 


Ethnic groups and languages 376 

Writing systems and literature 379 

The provinces and their rule 381 

Coinage 382 

Cities, architecture and art 383 






Christianity 414 

Indian religions (except Buddhism) 419 

Local religions 421 


Written sources 425 

Inscriptions in Buddhist complexes 432 

Manuscripts from East Turkestan and Central Asia 434 

East Turkestan - a major Buddhist centre 440 



The first Arab invasion of Iran 444 

The conquest of Iraq 445 



ISBN 978-92-3-103211-0 Contents CONTENTS 

The conquest of Khuzistan 447 

The battle of Nihavend 448 

The conquest of Seistan 449 

The conquest of Khurasan 450 


The first Arab incursions into Transoxania 452 

The beginning of the conquest of Transoxania 453 

The struggle of the peoples of Central Asia against the Umayyads 454 

The struggle of the peoples of Central Asia against the c Abbasids 458 


Reasons for the fall of the Sasanian Empire 461 

Relations between Muslims and non-Muslims in the first century A.H 462 

First steps towards Islamization 463 

Iranian regional administration in the conquered territories 464 

Tabaristan and Dailam 465 

Zabul, Kabul, Gandhara and Ghur 466 

The Arabs in Sind 467 

The survival of pre-Islamic civilization 467 


MAPS 488 


INDEX 545 

History of civilizations of Central Asia 558 


ISBN 978-92-3-103211-0 Contents TITLE 

History of Civilizations of Central Asia 

The crossroads of civilizations: 

a.d. 250 to 750 


Editor: B. A. Litvinsky 

Co-editors: Zhang Guang-da 

and R. Shabani Samghabadi 

Multiple History Series 
UNESCO Publishing 


ISBN 978-92-3-103211-0 Contents COPYRIGHT 

The authors are responsible for the choice and the presentation of the facts contained in 

this book and for the opinions expressed therein, which are not necessarily those of 

UNESCO and do not commit the Organization. 

The designations employed and the presentation of material throughout this publication 

do not imply the expression of any opinion whatsoever on the part of UNESCO 

concerning the legal status of any country, territory, city or area or of its authorities, or 

concerning the delimitation of its frontiers or boundaries. 

Compiled by I. Iskender-Mochiri 

Text revision by Dr D. W. MacDowall 

English text edited by Jana Gough 

Composed by Editions du Mouflon, 94270 Le Kremlin-Bicetre (France) 
Printed by Imprimerie Darantiere, 21800 Quetigny (France) 

Published in 1996 by the United Nations Educational, Scientific and Cultural 
Organization 7 place de Fontenoy, 75352 Paris 07 SP 

ISBN 92-3-103211-9 

© UNESCO 1996 



ISBN 978-92-3-103211-0 Contents PREFACE 


Federico Mayor 
Director-General of UNESCO 

One of the purposes of UNESCO, as proclaimed in its Constitution, is 'to develop and 
to increase the means of communication between. . . peoples and to employ these means 
for the purposes of mutual understanding and a truer and more perfect knowledge of each 
other's lives'. The History of the Scientific and Cultural Development of Mankind, pub- 
lished in 1968, was a major early response on the part of UNESCO to the task of enabling 
the peoples of the world to have a keener sense of their collective destiny by highlighting 
their individual contributions to the history of humanity. This universal history - itself now 
undergoing a fundamental revision - has been followed by a number of regional projects, 
including the General History of Africa and the planned volumes on Latin America, the 
Caribbean and on aspects of Islamic culture. The History of Civilizations of Central Asia 
is an integral part of this wider enterprise. 

It is appropriate that the second of UNESCO's regional histories should be concerned 
with Central Asia. For, like Africa, Central Asia is a region whose cultural heritage has 
tended to be excluded from the main focus of historical attention. Yet from time immemo- 
rial the area has served as the generator of population movements within the Eurasian 
land-mass. The history of the ancient and medieval worlds, in particular, was shaped to an 
important extent by the succession of peoples that arose out of the steppe, desert, oases and 
mountain ranges of this vast area extending from the Caspian Sea to the high plateaux of 
Mongolia. From the Cimmerians mentioned in Homer's Odyssey, the Scythians described 
by Herodotus, the Hsiung-nu whose incursions led the emperors of China to build the Great 
Wall, the sixth-century Turks who extended their empire to the boundaries of Byzantium, 
the Khitans who gave their name to ancient Cathay, through to the Mongols who erupted 
into world history in the thirteenth century under Genghis Khan, the nomadic horsemen 
of Central Asia helped to define the limits and test the mettle of the great civilizations of 
Europe and Asia. 



ISBN 978-92-3-103211-0 Contents PREFACE 

Nor is it sufficient to identify the peoples of Central Asia simply with nomadic cultures. 
This is to ignore the complex symbiosis within Central Asia itself between nomadism 
and settlement, between pastoralists and agriculturalists. It is to overlook above all the 
burgeoning of the great cities of Central Asia such as Samarkand, Bukhara and Khiva, 
which established themselves in the late Middle Ages as outstanding centres of intellectual 
inquiry and artistic creation. The seminal writings of the philosopher-scientist Avicenna 
(a native of Bukhara) and the timeless masterpieces of Timurid architecture epitomize the 
flowering of medieval culture in the steppes and deserts of Central Asia. 

The civilizations of Central Asia did not, of course, develop in a vacuum. The impact 
of Islam was pervasive and fundamental. The great civilizations on the periphery of the 
Eurasian continent likewise exerted an important influence on these lands. For some 1,500 
years this arid inland sea - far removed from the earth's true oceans - was crucial as the 
route along which merchandise (notably silk) and ideas flowed between China, India, Iran 
and Europe. The influence of Iran - although the core of its civilization lies in South-West 
Asia - was particularly strong, to the extent that it is sometimes difficult to establish a clear 
boundary between the civilization of the Iranian motherland and that of the outlying lands 
of Central Asia. 

To the rich variety of peoples of Central Asia was thus added a multiplicity of external 
influences. For century after century, the region experienced the influx of foreign art and 
ideas, colliding and merging with the indigenous patterns of Central Asia. Migrations and 
the recurrent shock of military invasion, mingling and displacing peoples and cultures, 
combined to maintain the vast region in flux. 

The systole and diastole of population movements down the ages add to the difficulty of 
delimiting a region whose topology alone does not prescribe clear boundaries. Thus, when, 
at the nineteenth session of its General Conference, UNESCO decided to embark on a His- 
tory of Civilizations of Central Asia the first problem to be resolved was to define the scope 
of the region concerned. Subsequently, at a UNESCO meeting held in 1978, it was agreed 
that the study on Central Asia should deal with the civilizations of Afghanistan, north- 
eastern Iran, Pakistan, northern India, western China, Mongolia and the former Soviet 
Central Asian republics. The appellation 'Central Asia', as employed in this History, refers 
to this area, which corresponds to a clearly discernible cultural and historical reality. 

UNESCO's call to specialists, and particularly to scholars native to the region, to partic- 
ipate in the undertaking met with a wide and generous response. The project was deemed 
by academics to be an excellent opportunity to draw back the curtain that had veiled 
Central Asia for so long. However, none were in any doubt as to the huge dimensions 
of the task. 



ISBN 978-92-3-103211-0 Contents PREFACE 

An ad hoc International Scientific Committee was formed in 1980 to plan and prepare 
the work, which it was agreed should cover, in six volumes, the history of Central Asia 
from earliest times to the present day. The Committee's initial task was to decide where 
pre-eminence should be given in the very wide canvas before it. In due course, a proper 
balance was struck and teams of editors and authors were selected. 

The preparation of the History of Civilizations of Central Asia is now well advanced. 
The best resources of research and archaeology have been used to make the work as thor- 
ough as possible, and countless annals consulted in major centres throughout the region. It 
is my sincere wish that this, the third volume, and those that follow will bring instruction 
and pleasure to readers all over the world. 

It remains for me to thank the President, Rapporteur and members of the International 
Scientific Committee, and the editors, authors and teams of specialists who have collabo- 
rated to shed new light on Central Asia with this detailed account of its vital and stirring 
past. I am sure it will prove a notable contribution to the study and mutual appreciation of 
the cultures that are the common heritage of mankind. 



ISBN 978-92-3-10321 1-0 Contents DESCRIPTION OF THE PROJECT 


M. S. Asimov 

The General Conference of UNESCO, at its nineteenth session (Nairobi, October, Novem- 
ber 1976), adopted the resolution which authorized the Director-General to undertake, 
among other activities aimed at promoting appreciation and respect for cultural identity, 
a new project on the preparation of a History of Civilizations of Central Asia. This project 
was a natural consequence of a pilot project on the study of Central Asia which was 
approved during the fourteenth session of the UNESCO General Conference in Novem- 
ber 1966. 

The purpose of this pilot project, as it was formulated in the UNESCO programme, was 
to make better known the civilizations of the peoples living in the regions of Central Asia 
through studies of their archaeology, history, languages and literature. At its initial stage, 
the participating Member States included Afghanistan, India, Iran, Pakistan and the former 
Soviet Union. Later, Mongolia and China joined the UNESCO Central Asian project, thus 
enlarging the area to cover the cultures of Mongolia and the western regions of China. 

In this work, Central Asia should be understood as a cultural entity developed in the 
course of the long history of civilizations of peoples of the region and the above delimita- 
tion should not be taken as rigid boundaries either now or in the future. 

In the absence of any existing survey of such large scope which could have served as 
a model, UNESCO has had to proceed by stages in this difficult task of presenting an 
integrated narrative of complex historical events from earliest times to the present day. 

The first stage was designed to obtain better knowledge of the civilizations of Central 
Asia by encouraging archaeological and historical research and the study of literature and 
the history of science. A new project was therefore launched to promote studies in five 
major domains: the archaeology and the history of the Kushan Empire, the history of the 
arts of Central Asia, the contribution of the peoples of Central Asia to the development of 
science, the history of ideas and philosophy, and the literatures of Central Asia. 

An International Association for the Study of Cultures of Central Asia (IASCCA), a 
non-governmental scholarly organization, was founded on the initiative of the Tajik scholar 



ISBN 978-92-3-10321 1-0 Contents DESCRIPTION OF THE PROJECT 

B. Gafurov in 1973, assembling scholars of the area for the co-ordination of interdis- 
ciplinary studies of their own cultures and the promotion of regional and international 

Created under the auspices of UNESCO, the new Association became, from the very 
beginning of its activity, the principal consultative body of UNESCO in the implementation 
of its programme on the study of Central Asian cultures and the preparation of a History 
of Civilizations of Central Asia. 

The second stage concentrated on the modern aspects of Central Asian civilizations and 
the eastward extension of the geographical boundaries of research in the new programme. 
A series of international scholarly conferences and symposia were organized in the coun- 
tries of the area to promote studies on Central Asian cultures. 

Two meetings of experts, held in 1978 and 1979 at UNESCO Headquarters, concluded 
that the project launched in 1967 for the study of cultures of Central Asia had led to con- 
siderable progress in research and contributed to strengthening existing institutions in the 
countries of the region. The experts consequently advised the Secretariat on the methodol- 
ogy and the preparation of the History. On the basis of its recommendations it was decided 
that this publication should consist of six volumes covering chronologically the whole his- 
tory of Central Asian civilizations ranging from their very inception up to the present. 
Furthermore, the experts recommended that the experience acquired by UNESCO during 
the preparation of the History of Scientific and Cultural Development of Mankind and of 
the General History of Africa should also be taken into account by those responsible for 
the drafting of the History. As to its presentation, they supported the opinion expressed 
by the UNESCO Secretariat that the publication, while being a scholarly work, should be 
accessible to a general readership. 

Since history constitutes an uninterrupted sequence of events, it was decided not to give 
undue emphasis to any specific date. Events preceding or subsequent to those indicated 
here are dealt with in each volume whenever their inclusion is justified by the requirements 
of scholarship. 

The third and final stage consisted of setting up in August 1980 an International Scien- 
tific Committee of nineteen members, who sit in a personal capacity, to take reponsibility 
for the preparation of the History. The Committee thus created included two scholars from 
each of the seven Central Asian countries - Afghanistan, China, India, Islamic Republic of 
Iran, Pakistan, Mongolia and the former USSR - and five experts from other countries - 
Hungary, Japan, Turkey, the United Kingdom and the United States of America. 

The Committee's first session was held at UNESCO Headquarters in December 1980. 
Real work on the preparation of the publication of the History of Civilizations of Central 



ISBN 978-92-3-10321 1-0 Contents DESCRIPTION OF THE PROJECT 

Asia started, in fact, in 1981. It was decided that scholars selected by virtue of their quali- 
fications and achievements relating to Central Asian history and culture should ensure the 
objective presentation, and also the high scientific and intellectual standard, of this History. 

Members of the International Scientific Committee decided that the new project should 
correspond to the noble aims and principles of UNESCO and thereby should contribute 
to the promotion of mutual understanding and peace between nations. The Committee 
followed the recommendation of the experts delineating for the purpose of this work the 
geographical area of Central Asia to reflect the common historical and cultural experience. 

The first session of the International Committee decided most of the principal matters 
concerning the implementation of this complex project, beginning with the drafting of 
plans and defining the objectives and methods of work of the Committee itself. 

The Bureau of the International Scientific Committee consists of a president, four vice- 
presidents and a rapporteur. The Bureau's task is to supervise the execution of the project 
between the sessions of the International Scientific Committee. The reading committee, 
consisting of four members, was created in 1986 to revise and finalize the manuscripts 
after editing Volumes I and II. Another reading committee was constituted in 1989 for 
Volumes III and IV 

The authors and editors are scholars from the present twelve countries of Central Asia 
and experts from other regions. Thus, this work is the result of the regional and of the inter- 
national collaboration of scholars within the framework of the programme of the United 
Nations Educational, Scientific and Cultural Organization (UNESCO). 

The International Scientific Committee and myself express particular gratitude to Mrs 
Irene Iskender-Mochiri for her arduous and selfless work in preparing the first three vol- 
umes for the press. 

It is our sincere hope that the publication of the third volume of the History of Civiliza- 
tions of Central Asia will be a further step towards the promotion of the cultural identity 
of the peoples of Central Asia, strengthening their common cultural heritage and, conse- 
quently, will foster a better understanding among the peoples of the world. 



ISBN 978-92-3-103211-0 



SCIENTIFIC COMMITTEE (in alphabetical 


Dr F. R. Allchin (United Kingdom) 

Professor M.S. Asimov (Tajikistan) 


Editor of Volume IV (Parts I and II) 

Dr N. A. Baloch (Pakistan) 

Professor M. Bastani Parizi (Islamic 
Republic of Iran) 

Professor S. Bira (Mongolia) 

Professor A. H. Dani (Pakistan) 
Editor of Volume I 

t Professor K. Enoki (Japan) 

Professor G. F. Etemadi (Afghanistan) 
Co-editor of Volume II 

Professor J. Harmatta (Hungary) 
Editor of Volume II 

Professor Liu Cunkuan (People's 
Republic of China) 

Dr L. I. Miroshnikov (Russian Federation) 

Professor S. Natsagdorj (Mongolia) 

Professor B. N. Puri (India) 
Co-editor of Volume II 

Professor M. Ft. Z. Safi (Afghanistan) 

Professor A. Sayili (Turkey) 

Dr R. Shabani Samghabadi (Islamic 
Republic of Iran) 
Co-editor of Volume III 

Professor D. Sinor (United States of 

t Professor B. K. Thapar (India) 

Professor Zhang Guang-Da (People's 
Republic of China) 
Co-editor of Volume III 



ISBN 978-92-3-103211-0 




Professor C. Adle 
(Islamic Republic of Iran) 
Editor of Volume V 

Professor D. Alimova 

Professor M. Annanepesov 

Professor M.S. Asimov 


President and Editor of Volume IV 

(Parts I and II) 

Professor K. Baipakov 
Co-editor of Volume V 

Professor S. Bira 

Professor A. H. Dani 


Editor of Volume I 

Professor H.-P Francfort 


Professor I. Habib 


Editor of Volume V 

Dr L. Miroshnikov 
(Russian Federation) 

Professor D. Sinor 
(United States of America) 

Dr A. Tabyshalieva 
(Kyrgyz Republic) 

Professor I. Togan 

Professor H. Umemura 

Professor Wu Yungui 
(People's Republic of China) 



ISBN 978-92-3-103211-0 Contents COMMITTEE 


Professor A. D. H. Bivar Professor J. Harmatta 

(United Kingdom) (Hungary) 

Professor R. N. Frye Professor D. Sinor 

(United States of America) (United States of America) 



ISBN 978-92-3-103211-0 




t P. G. Bulgakov 

K. Chakrabarti 
Centre for Historical Studies 
Jawaharlal Nehru University 
New Delhi 110067, India 

N. N. Chegini 
Iran Bastan Museum 
Khiaban-e Imam Khomeini 
Tehran 11364/9364 
Islamic Republic of Iran 

A. H. Dani 


Centre for the Study of Civilizations of 

Central Asia 

Quaid-i-Azam University 

Islamabad, Pakistan 

PH. Gignoux 

Ecole Pratique des Hautes Etudes 

Section des Sciences religieuses 


45, rue des Ecoles 

75005 Paris, France 

J. Harmatta 
Budapest 1015, Hungary 

A. H. Jalilov 
Tajik State University 
Dushanbe, Tajikistan 

t A. L. Khromov 

S. G. Klyashtorny 
Institute of Oriental Studies 
Sector of Turco-Mongol Studies 
DvortsovayaNab., 18 
St Petersburg, 
Russian Federation 

A. I. Kolesnikov 
Institute of Oriental Studies 
DvortsovayaNab., 18 

St Petersburg, Russian Federation 

L. R. Kyzlasov 
Faculty of History 
Moscow State University 
Moscow, Russian Federation 



Institute of Oriental Studies 

Rojdesvenko Street 12 

Moscow 103753, Russian Federation 



ISBN 978-92-3-103211-0 



B. I. Marshak 
State Hermitage Museum 
St Petersburg, 
Russian Federation 

Mu Shun-Ying 

Archaeological Institute of Xinjiang 
Academy of Social Sciences 
Urumqi, People's Republic of China 

N. N. Negmatov 
Institute of History 
Tajik Academy of Sciences 
Dushanbe, Tajikistan 

E. E. Nerazik 

Institute of Ethnology and Anthropology 

Leninskiy prospekt, 32A 

Moscow 117334, Russian Federation 

A. V. Nikitin 

The State Hermitage Museum 

St Peterburg, Russian Federation 

D. Sinor 

Distinguished Professor Emeritus of Uralic 

and Altaic Studies 

Indiana University 

Goodbody Hall 157 

Bloomington, Indiana 47405-2401 

United States of America 

A. Tafazzoli 

Faculty of Letters and Human Sciences 

Tehran, Islamic Republic of Iran 

M. I. Vorobyova-Desyatovskaya 
Institute of Oriental Studies 
DvortsovayaNab., 18 
St Petersburg, Russian Federation 

Wang Yao 

Central Institute for National Minorities 
Department of Tibetan Studies 
Beijing, People's Republic of China 

M. H. Zamir Safi 
Department of History 
Faculty of Social Sciences 
Kabul University 
Kabul, Afghanistan 

E. V. Zeimal 

Oriental Department 

State Hermitage Museum 

St Petersburg, Russian Federation 

Zhang Guang-Da 
Department of History 
Beijing University 
Beijing 100871 
People's Republic of China 



ISBN 978-92-3-103211-0 


Other collaborating specialists 

Other collaborating specialists 

M. E. Bastani Parizi 

Faculty of Letters and Human Sciences 

University of Tehran 

Tehran, Islamic Republic of Iran 

G. S. Humayun 
Faculty of Literature 
Kabul University 
Kabul, Afghanistan 

M. Hussain Shah 
Faculty of Social Sciences 
Kabul University 
Kabul, Afghanistan 

M. A. Joyenda 

Afghan Institute of Archaeology 
International Centre for Kushan Studies 
Kabul, Afghanistan 

A. Khodadadian 

Faculty of Letters and Human Sciences 
Tehran, Islamic Republic of Iran 

B. N. Puri 
B-58, Sector A 
Lucknow 226006, India 

A. Rahman 

Department of Archaeology 
University of Peshawar 
Peshawar, Pakistan 



ISBN 978-92-3-103211-0 Contents HISTORICAL INTRODUCTION 


B. A. Litvinsky and Zhang Guang-da 


The Sasanians 24 

The Guptas 26 

The Sui dynasty 27 

The T'ang dynasty 28 

Oasis states 29 

Ecology, geography and climate 29 

Nomadic societies 31 

The Hsiung-nu and the Huns 33 

The Silk Route 35 

Cross-cultural influences 37 

The end of the period covered by Volume II of the History of Civilizations of Central Asia 
saw the weakening and collapse of the powerful Parthian, Han and Kushan empires. The 
present volume deals with the period c. a.d. 250-750, which witnessed the rise of mighty 
new empires (Sasanian, Gupta, Sui and T'ang; and the Arab caliphate) on the fringes of 
Central Asia Central Asia. It also saw the successive movements of nomadic peoples (the 
Huns, Alan tribes, Chionites, Kidarites and Hephthalites, Turks, Tiirgesh, Karluks, Uighurs 
and other Turkic tribal confederations) that played a major and at times decisive role in the 
later ethnic and political history of the region. 

See Map 1. 



ISBN 978-92-3-103211-0 Contents The Sasanians 

The Sasanians 

In Iran the entire period is covered by the Sasanian Empire (a.d. 224-651), the successor 
to the Parthian Empire (c. 250 B.C. - a.d. 224). The end of the first century and the begin- 
ning of the second were characterized by chronic internal disturbances in the Parthian state, 
which was also weakened by the constant wars with Rome. The Parthians were particularly 
unsuccessful in their struggle against the Roman emperor Trajan in the years 113-117. 
Although they later managed to regain their lost lands and achieve a certain degree of 
stability, the wars with Rome proved highly exhausting. 1 Entire provinces, including Hyr- 
cania, were lost and Margiana became independent. 

The beginning of the third century was marked by rivalry between the Parthian king 
Vologases V (207/8-221/2) and his brother Ardavan V (c. 213- 224), who ruled indepen- 
dently, as well as by further battles against the Romans. It was at that time that a new 
dynasty emerged: the Sasanians. According to Lukonin: 

The emergence of the Sasanians and the organization of a new state in Iran in the first quarter 
of the third century a.d. meant not only a change of dynasty. These developments were due 
to deep-seated economic and political factors. The growth of commodity production brought 
about mainly by the exploitation of the peasantry and slaves, the growth in demand for crafts 
and agricultural produce linked to the revival of trade routes running through Iran to China 
and India, and the general crisis in the system of slave-ownership which affected the Mediter- 
ranean in the third century and reached Iran were all expressions of a new stage in the history 
of that country which created an urgent need for new organizational and political forms. 
Ardashir [226- 241], the son of Papak and scion of a local national dynasty, supported by a 
broad section of the increasingly feudal nobility both priestly and military, rapidly managed 
to unite scattered princedoms and domains around Persia (modern Fars province), an area 
associated from earliest times with the national unity of the entire country. 2 

As early as the third century the Sasanians destroyed the large Kushan Empire and annexed 
a part of its domains, including Bactria (Tokharistan). This region subsequently became 
part of the small Kushano-Sasanian kingdom and then, at times, part of the vast and mighty 
Sasanian Empire itself, which stretched from the southern regions of western Central Asia 
and Afghanistan to the Transcaucasus, Mesopotamia and part of Arabia. To the west, the 
Sasanians shared a border with the Roman Empire and subsequently with Byzantium, and 
this was the theatre of almost constant wars. Similar wars were waged on its eastern border 
with the Later Kushans, the Chionites, the Hephthalites and the Turks. Some of these tribes 
invaded the territory of Iran and their notables played a part in domestic Iranian political 

1 Lepper, 1948. 

2 Lukonin, 1961, p. 5. 



ISBN 978-92-3-103211-0 Contents The Sasanians 

The Sasanians had a developed bureaucratic and military system as well as a com- 
plex administrative, social and Zoroastrian priestly hierarchy. The reforms of Khusrau I 
Anushirvan (531-579) marked not only the establishment of a well-balanced, comprehen- 
sive and fairly centralized system of administration but also the completion of one stage in 
the feudalization of society. 

A vigorous culture flourished in the Sasanian Empire. Many works of religious, artistic 
and scientific literature were produced and there was a large body of secular literature such 
as historical epics and poems, legal, geographic and other works, including some in poetic 
form. There were many translations from other languages, including Greek, Syriac and 
Indian languages. Majestic works of architecture have been preserved, which testify to the 
engineering and architectural genius of their creators. Among the arts represented were 
sculpture, bas-relief, painting, toreutics, glyptic and representations on medals and coins. 
It was a new and highly progressive stage in the development of the art of the East. 

Art, crafts and architecture bear traces of previous Iranian, as well as Roman, Byzantine, 
Transcaucasian and Central Asian artistic creation. Sasanian spiritual and material life and 
culture also exerted a considerable influence on those of many neighbouring and more 
distant peoples. Sasanian works of art and crafts were to be found from France to China and 
Japan and many examples have been discovered in the Urals and Siberia. The traditions, 
motifs and rites created under the Sasanians continued into the Islamic period. 3 

The Sasanian shahanshah (king of kings) of Iran was the bearer of kingly khwarnah 
(Avestan, khvarenah; Old Persian, farnah; New Persian, farr), in other words, he embodied 
the happiness and destiny of the royal dynasty and the entire state. 4 The Sasanian monarch 
was considered the earthly representative and counterpart both of the Zoroastrian religion 
and of its supreme creator deity, Ohrmazd (Avestan, Ahura Mazda). Consequently, the king 
in Zoroastrian belief, as sovereign of Iran and the entire corporeal world, was the divinely 
designated protector, religious and secular authority, and guide of the material creation. 
He served, in the corporeal, the same roles as does Ohrmazd as universal sovereign of 
both the material and spiritual worlds. 5 Identical or similar conceptions connected with 
sacral kingship were found in India and China, although both countries enjoyed a degree 
of religious tolerance that did not exist in Iran. 

Zoroastrianism was the prevailing faith (and also the state religion) in Iran, with Chris- 
tianity and Judaism existing alongside it. A syncretic Manichaean religion arose in the 
Mesopotamian part of the empire. From Iran, Zoroastrianism, Nestorianism, Judaism and 

3 The Cambridge History of Iran, 1983, Vol. 3, Parts 1 and 2; Christensen, 1944. 

4 Bailey, 1971. 

5 Choksy, 1988; see also Widengren, 1959; Frye, 1984. 



ISBN 978-92-3-103211-0 Contents The Guptas 

Manichaeism spread east to western Central Asia, Afghanistan, East Turkestan, Mongolia 
and China. 

Like the Parthian Empire, the Kushan Empire began to decline in the third century and 
severe blows were dealt to it by the Sasanians. The Indian part of the Kushan Empire also 
declined, losing its influence first in the Ganges valley. The Yaudheya Republic (Yaudheya- 
gana on coin inscriptions) - situated on the plain between the Sutlej and the Jamuna and in 
northern Rajasthan - played a major part in the struggle against the Kushans. In the second 
century, western India was controlled by the dynasty of the Western Satraps, who extended 
their territories, but were eventually incorporated in the Gupta Empire by Chandragupta II 
at the end of the fourth century. 

The Guptas 

The region of Magadha rose to prominence during the third century largely because of its 
situation on the lower reaches of the Ganges close to the shores of the Bay of Bengal. At the 
beginning of the fourth century it became the political centre of the Gupta Empire, which 
rapidly united most of northern India. Chandragupta I (c. 319-335 or 350) played a major 
role in this, and after his death he was given the splendid title of 'Great King of Kings'. His 
son, Samudragupta (c. 350-c. 375), was a skilful politician, a bold and successful military 
commander and a patron of the arts and sciences. It was during his rule that the Gupta 
Empire took shape in the valley of the Ganges. Its nucleus was surrounded by a belt of 
territories dependent to a greater or lesser extent on the Guptas and Samudragupta even led 
a campaign deep into southern India. 

The territory of the Guptas reached its greatest extent under Chandragupta II (c. 375-c. 
415) after he defeated the Western Satraps - a period known as the 'Gupta Golden Age'. 
There are even references to a campaign against Bactria. But the Gupta Empire weakened 
and eventually broke up under the attacks of new invaders, among whom the Hephthalites 
- the Hunas of Indian sources - played an important role. Many regions of northern India 
fell under the control of Turkic invaders, until Harsha of Kanauj (606-647), a powerful 
conqueror, rebuilt a strong empire in the north. 

The Guptas played a major role in the history and culture of India and of the neigh- 
bouring countries. Although the empire had no standard administrative divisions or unified 
administrative system, there was a marked increase in the level of centralization. Record- 
keeping was highly developed and there was a complex system for the registration of land, 



ISBN 978-92-3-103211-0 Contents The Sui dynasty 

donations, income and expenditure. Agriculture was developed, as were crafts, building, 
and foreign and domestic trade. 6 According to Rowland: 

Seldom in the history of peoples do we find a period in which the national genius is so fully 
and typically expressed in all the arts as in Gupta India. Here was florescence and fulfilment 
after a long period of gradual development, a like sophistication and complete assurance in 
expression in music, literature, the drama, the plastic arts and architecture. The Gupta period 
may well be described as 'classic' in the sense of the word describing a norm or degree 
of perfection never achieved before or since, and in the perfect balance and harmony of all 
elements, stylistic and iconographic elements inseparable in importance. 

Sanskrit became the official language of the Gupta court. The great Indian epic, the 
Mahabharata, underwent a final recension as a document of a unified India under a godly 
Imperial race; the Ramayana enjoyed a renewed popularity. . . It was in this period that the 
Indian theatre, which, just like western drama, traced its origins to the performance of church 
spectacles or miracle plays, reached the extraordinary perfection of dramatic structure and 
richness of metaphor that characterize the 'Toy cart' and the famed Kalidasa's rich and sen- 
suous poetic drama 'Sakuntala'. 7 

Buddhism ( Mahayana and Hinayana) flourished in Gupta India but external and inter- 
nal factors gradually contributed to its decline. Hinduism was regenerated and absorbed 
many Buddhist beliefs. Gupta India contained many universally known centres of erudi- 
tion, including the monastery at Nalanda. From Gupta India, ideas, teaching, scientific 
discoveries and also miscellaneous goods, works of art and literature, scientific and reli- 
gious writings, preachers, merchants and craftsmen spread throughout Central Asia to the 
lands of the southern seas, the Mediterranean and East Asia. 

The Sui dynasty 

In China the end of the Han dynasty (206 B.C. - a.d. 220) was a dark period. Rebellions 
and internecine strife devastated the flourishing central regions of the Han Empire; towns 
lay in ruins, the fields were empty and trade routes ceased to operate. In 220 a new dynasty, 
the Wei, took power. During its rule the processes of disintegration continued apace and a 
mere 45 years later another dynasty, the Chin, was established. A short period of unity was 
followed by fresh rebellions, internecine wars and the movement of northern and north- 
western nomads, proto-Turkic, Tungusic and tribes of Tibetan-Chiang stock, who had been 
fighting among themselves for years. Separate states were formed in the north and south 
of China. 

6 The Age of Imperial Unity, 1951; The Classical Age, 1954; Altekar and Majumdar (eds), 1946; Banerji, 
1933; Dandekar, 1941; Mookerji, 1952; Bongard-Levin and Il'in, 1969a. 

7 Rowland, 1970, pp. 215-16. 



ISBN 978-92-3-103211-0 Contents The Tang dynasty 

In 581 the Sui dynasty was founded: four centuries of rebellion and division gave way 
to a period of unity and centralization. Strong measures were taken to set up a bureaucratic 
type of government and major building works were initiated. The Sui emperors started 
wars on the borders of the empire but obtained few successes. In comparison with the 
advanced consolidation of the Han period, the state's territory during all the following 
periods (including that of the Sui) was significantly smaller. China's influence in Central 
Asia declined gradually, leaving few traces; trade routes also declined and cultural ties 
were weakened. 

The T'ang dynasty 

In 618 the T'ang dynasty came to power in China. Its real founder was the emperor T'ai- 
tsung (627-649), during whose 23-year reign many laws were codified and all aspects 
of life were carefully regulated. T'ang capitals and other cities flourished, as did trade. 
Although the empire reached its zenith under Hsiian-tsung (713-756), his reign also marked 
the start of the dynasty's decline. 

China's consolidation and development during the Early T'ang period led to an expan- 
sion of its territory, especially to the west, where practically all East Turkestan was incor- 
porated into the empire. T'ang troops advanced to the borders of western Central Asia, but 
were unable to consolidate their hold. In the last quarter of the seventh century, the T'ang 
Empire started to clash in East Turkestan with the Tibetans, who had become very power- 
ful. Tibetan expansion reached its zenith in the middle and the third quarter of the eighth 
century, when they held a large area of East Turkestan. 

The T'ang maintained large military forces in the north to fight the Turks, who had 
united under the First (552-630) and the Second (682-744) Turk Kaghanates. The year 
744 saw the establishment of the Uighur Kaghanate, which rapidly became very power- 
ful. After an interlude, the Uighurs began to play a role in internal Chinese affairs at the 
invitation of the T'ang government. 

The process of urbanization continued, accompanied by an expansion in crafts, and in 
domestic and foreign trade. In the area of spiritual life, Buddhism gained in importance. 
Outstanding works of literature, architecture and art were created. According to Schafer: 

How the Western Regions contributed to China and then T'ang China contributed her arts and 
manners to her neighbours of the medieval Far East, especially to Japan, Korea, Turkestan, 
Tibet and Annam, is a rather well-known story. To mention the arts of xylography, city plan- 
ning, costume design, and versification is only to hint at the magnitude of the cultural debt 
which these peripheral countries owed to T'ang. We are also familiar with the material goods 
sought by foreigners in China or taken abroad by the Chinese themselves: luxuries like silk 



ISBN 978-92-3-10321 1-0 Contents Ecology, geography and climate 

textiles, wine, ceramics, metalwork, and medicines, as well as such minor dainties as peaches, 
honey and pine nuts, and, of course, the instruments of civilization, great books and fine 
paintings. 8 

In summary, it is clear that the lands bordering on Central Asia (especially Iran and China) 
acted as mighty generators of military and political power. On their territories, great cen- 
tralized states were established, mighty empires which played a crucial role in the fate of 
the various peoples of Central Asia and in diplomatic and economic history. The Indian, 
Iranian and Chinese civilizations also played an outstanding part in the development of the 
civilization of Central Asia as a whole, contributing to material culture, armaments and the 
design and construction of cities as well as to science, philosophy, literature and religion. 

Oasis states 

In the middle of Central Asia lay oasis states with settled agricultural (and also, to some 
extent, nomadic) populations and a developed urban life. They included Tokharistan (Bac- 
tria), Margiana, Sogdiana, Khwarizm, and many small kingdoms in the Tarim basin such 
as Kucha and Khotan, all of which enjoyed several common features - we shall mention 
just a few of them. The character of these states was monarchic in principle, but a theo- 
cratic monarchy (or something resembling it in the case of Khotan, for example) had been 
established in some of them. There was a secular, hierarchical system in which vassalage 
was developing, and it is possible to speak of the growth of feudalism. In all these states, 
the period was characterized by the expansion of productive forces and by a complex sys- 
tem of trade. The Sogdians were especially renowned as traders from Byzantium to China 
- they not only travelled back and forth in caravans but founded entire cities 9 and acted 
as the transmitters of cultural values. The oasis states generated a very high level of urban 
culture, as testified by the creation of outstanding works of art which became part of the 
Eastern heritage. 

Ecology, geography and climate 

The ecological conditions prevailing in Central Asia are as follows. Although its territory 
lies on the same latitude as Spain and the southern half of France, the natural conditions are 
quite different. It is a land of boundless sandy deserts. There are mountains rising above 
the clouds, covered in eternal snows, and large mountain glaciers over alpine meadows. 
Swollen rivers flow from these glaciers and snowfields, rushing down into the plains with 

8 Schafer, 1963, p. 2; see also Wright, 1959; Kryukov et al., 1979; 1984. 

9 Chugevskiy, 1971; Pulleyblank, 1952. 



ISBN 978-92-3-10321 1-0 Contents Ecology, geography and climate 

their cultivated lands, villages and towns. From the foot of the mountains, the deserts of 
western Central Asia - the Kyzyl Kum, the Kara Kum and the Ustyurt - stretch for more 
than 1,000 km to the Caspian Sea. Middle Asia (a region of the former Soviet Central 
Asia) lies in the internally drained part of Eurasia, that is in the Aral-Caspian (Turanian) 
plain, where the above-mentioned deserts and the T'ien Shan and Pamir- Alai (the highest 
mountain systems of this region) are located. Further east, regions with similar landscapes 
stretch all the way to Mongolia. 

The main climatic features throughout most of western Central Asia are an abundance 
of heat and light, aridity and a continental pattern. The variations in daily and monthly 
temperatures can be as great as 40 ° C. On the plains, the frost-free period may last for 
up to 250 days. The summer is long, oppressive and dry. In the far south the temperature 
reaches 50 ° C in the shade. The lowlands are among the driest regions of Eurasia; thus the 
precipitation in the Taklamakan desert is insignificant - 10-15 mm per annum. In Dzun- 
garia, the precipitation varies between 200 mm in the northern foothills and 400 mm in the 
northern plains. In the Alai and T'ien Shan mountains the climate is much more humid, and 
precipitation rises to 500-700 mm per annum. However, its volume does not automatically 
correspond to the relief and it is very low on some high mountain plateaux: for example, 
60-80 mm per annum in the eastern Pamirs. 

Types of vegetation associated with very arid conditions are typical of the deserts of 
Central Asia: ephemeral or fast-ripening, their rapid development mainly occurs during the 
humid spring. The nature of the desert vegetation depends on the ecological environment, 
such as the nature of the soil, its chemical composition, and humidity. There are several 
types of desert vegetation; thus a given region may be used only at a particular time of the 
year. Yet most desert pastureland can be used all the year round. Its productivity is low, 
and one sheep requires 5-10 ha of pasture. 

The vegetation in the mountains and on the plains is much richer. Here there are separate 
spring, summer, autumn and winter pastures. On the plains, vegetation begins to grow in 
early spring and is already withering by the end of May. Livestock graze here in the spring. 
These pastures are also used later on, but mainly for the exploitation of dried grass. From 
450 m to 1,500 m above sea level, mixed grasses and short-lived plants grow from April 
to the middle of June, after which they wither. These are good spring and early summer 
pastures. Higher up the humidity increases. The most productive summer pastures are to be 
found at that level, and also large areas of deciduous forest, including fruit-bearing trees. 
Higher still are sub-alpine and alpine meadows where climatic conditions permit livestock 
to graze for only three to three and a half summer months. These meadows are nevertheless 
held in particular esteem by herdsmen because of the high nutritive value of the vegetation. 



ISBN 978-92-3-103211-0 Contents Nomadic societies 

Ecology requires pastures to be used in rotation, since there are both flatlands and moun- 
tains. The cattle 'follow the spring', grazing first in the valley and then being moved higher, 
until they eventually reach the alpine meadows. In the autumn the cattle gradually move 
down, lingering at spots where fodder is available. Since winter fodder contains only half 
the nutrition of spring fodder, winter is a critical period in the annual cycle. 

Nomadic livestock-rearing also depends on the water supply, which differs from one 
area to another. In the mountains, humidity tends to be higher and there is sufficient water 
for animals. A completely different situation applies in desert areas and, to some extent, 
on the plains where there has always been a severe shortage of water. In both types of 
terrain, groundwater is widely used as well as water from rivers. In desert regions, water is 
obtained from wells 10-100 m deep. 

Needless to say, the foregoing is an extremely general description. Every part of the vast 
region of Central Asia has its own particular, and often fundamentally different, features. 

Nomadic societies 

The nomads who occupied the vast deserts, steppes and mountains were a very important 
factor in the history of Central Asia. The Han shu, the official history of the Former Han 
dynasty, describes the Wu-sun people who lived in Central Asia in ancient times: 'The land 
is covered in vegetation and is flat. . . [The people] do not work at cultivating the fields or 
planting trees, but in company with their stock animals they go in the same way as that of 
the Hsiungnu.' 10 Independent reports from the most distant areas of the populated world 
all agree that these were nomadic peoples. 11 

The Han shu's brief description conveys the essence of nomadism in Eurasia: its main 
ingredients were the migratory way of life and the operation of an economy in conditions 
which necessarily included vast expanses of pastureland. Nomadism implies the existence 
of a social entity developed for the operation of a specific economy, and a corresponding 
ecological niche where this entity can establish itself. It was only the combination of these 
two factors, ecological and social, that enabled nomadism to emerge and develop in various 
parts of Central Asia and of Eurasia as a whole. 

10 Hulsewe and Loewe, 1979, p. 144. Compare this with Strabo's description of European nomads in the 
Scythian orbit: 'As for the nomads, their tents, made of felt, are fastened on the wagons in which they spend 
their lives; and round about the tents are the herds which afford the milk, cheese, and meat on which they 
live; and they follow the grazing herds, from time to time moving to other places that have grass, living only 
in the march-meadows about Lake Macotis in winter, but also in the plains in summer' (Strabo, 7.3.17). 

11 Chinese sources on the Hsiung-nu (Taskin, 1968; 1973); on the Turks (Chavannes, 1903; Liu, 1958). 



ISBN 978-92-3-103211-0 Contents Nomadic societies 

There were several patterns of nomadism in Central Asia during the period from the 
third to the eighth century (as in the nineteenth and twentieth centuries). The most preva- 
lent form was longitudinal, meaning migration from north to south and vice versa. Lati- 
tudinal migration involves movement to east or west, conditioned by climatic variations, 
whereas vertical migration exploits the differences in altitude. Radial movements are those 
distributed around a central area according to the availability of pasture. Depending on 
the terrain, the length of migratory routes varied between 5-10 and 1,000 km. Weakened 
by lack of fodder during the winter, livestock moved slowly. Possessions were loaded on 
camels, while horses and unburdened camels followed at a distance, with sheep and other 
livestock moving close to the migrating group. Such groups travelled 15-25 km in the 
course of a day. 12 In mountainous regions there was an intricate and well-developed sys- 
tem of horizontal and vertical movement. 

The historical and archaeological evidence on nomads and their societies is supple- 
mented by ethnographic material. 13 Pure nomadism was not encountered often, even in 
Antiquity; neither was livestock-breeding the sole form of economic activity. Alongside 
the nomadic way of life, several types of seminomadic existence were widespread in areas 
with various forms of (partial) settlement and a fairly developed, settled economy based on 
agriculture and crafts. Nomads started to adopt a settled and sometimes even an urban way 
of life. 

Nomads were an important and integral part of the society of Eurasia. They established a 
characteristic economy, with highly developed techniques of livestock-rearing, the grazing 
of herds, nomadic movements and various forms of crafts and warfare. They also had 
a unique and highly developed social structure (based on the dual system of individual 
ownership of herds and communal ownership of pastureland), shamanistic religions and 
fine folk poetry, especially epics. Nomadic society is mobile in two senses: first, the internal 
mobility of the society itself; and, second, its external mobility, the capacity to execute 
rapid, far-reaching movements, both peaceful and military. 

Nomads usually lived close to oases, with their settled, urban, agricultural way of life, 
and various types of relationship existed between the two civilizations. According to Lat- 

In the pre-industrial age, the advanced urban-agricultural civilizations produced a great many 
things which the nomads wanted, but the nomads produced not nearly so much of what the 

12 Rudenko, 1961; Shakhmatov, 1963; Zhdanko, 1968; Tolybekov, 1971; Markov, 1976; Khozyaystvo kaza- 
khov, 1980. 

13 In addition to the sources listed, see also Vladimirtsov, 1948; Radloff, 1893; Barrou et al., 1973; Pastoral 
Production and Society, 1979; Khazanov, 1984; RoV kochevykh narodov v tsivilizatsii Tsentral'noy Azii, 
1974; Bongard-Levin and IT in, 1969b. 



ISBN 978-92-3-103211-0 Contents The Hsiung-nu and the Huns 

settled people wanted. When, therefore, the nomads were not rich enough to buy all that they 
wanted, but felt militarily strong, it was a great temptation to threaten, raid and even make 
deep invasions of the settled lands. (The 'imbalance' between the economies of China and 
the territories beyond the Great Wall was always very striking, but I believe that in the Indo- 
Iranian, Afghan and Arab lands, and the land of the Turkish-speaking peoples of Central 
Asia, both oasis dwellers and nomads. . . [there] may have been a much greater degree of 
integration.) 14 

The nomads often lived near settled oases or even within them, leading to a wide range 
of ethno-cultural and socio-economic contacts. Economic ties then became so close that 
at times it is possible to speak of the emergence of a common economic system with two 
intimately linked sectors, the settled and the nomadic. 

The entire period between 250 and 750 was characterized by major movements of 
nomadic tribes and peoples. What was the cause of these movements? At present we are 
unable to supply anything like a satisfactory answer. Tempting preliminary explanations 
such as changes in the climate cannot explain why Mongolia and the adjacent regions 
released wave after wave of nomads from the end of the first century B.C. up to the time 
of Chinggis Khan. These waves spread south to China, south-west to western Central Asia 
and East Turkestan and west to the Volga, the Black Sea and beyond that to Italy and 

The Hsiung-nu and the Huns 

The Hsiung-nu not only developed their society in Mongolia and the neighbouring regions 
of southern Siberia over the centuries but carried out raids and military expeditions in 
various directions. One of these directions was west: here they intervened in the struggle for 
East Turkestan and at the beginning of the second century B.C. the agricultural population 
of that region fell under their sway. The Hsiung-nu first engaged in a lengthy and critical 
struggle against other nomadic tribes, particularly the Wu-sun and the Yiieh-chih, and also 
with China and local settled agricultural princedoms. 

In the year 22 B.C. a catastrophic drought occurred. Many nomads then moved south, 
recognizing the authority of the Han, while others, the Northern Hsiung-nu (or Huns), 
moved north-west. Nomads began to take a more active part in the affairs of East Turkestan 
and penetrated the north-eastern corner of western Central Asia. The Hsiung-nu became 
even more active in the first century a.d. Their westward movement towards the Volga and 
the Don appears to have started in the first or second century a.d., when they had to pass 
through territory densely populated by Sarmatians and other tribes. On the lower reaches 

14 Lattimore, 1974, pp. 172-3. 



ISBN 978-92-3-103211-0 Contents The Hsiung-nu and the Huns 

of the Don they clashed with Alan tribes in the years 370-380. Referring to them as Huns, 
the Roman historian Ammianus Marcellinus (XXXI, 3.1) describes how 'the Huns killed 
and plundered [these tribes] and joined the survivors to themselves in a treaty of alliance'. 

It is clear that the Huns - as these tribes were called in the west - mixed with the Alans, 
and before that with the Sarmatians and later with the Goths of Ermenrichus. Then they 
invaded Pannonia and moved on into western Europe. There followed the age of Attila 
(434-453) and the battle of the Catalaunian Plains. 

The Huns were nomadic horsemen who played an important part in the 'great migration 
of peoples'. 15 According to Kryukov et al: 

A direct shock producing a kind of 'chain reaction': such was the westward migration of the 
branch of the Asiatic Hsiung-nu who left their original homeland in the second century and 
two and a half centuries later entered European history under the name of the Huns. But the 
'great migration of peoples' embraced not only Europe in the third to the sixth century. In East 
Asia a process began in the third century which was extremely similar to the one observed at 
the same time on the borders of the Roman Empire. A branch of the Hsiung-nu, the Hsien-pi, 
the Ti, the Tibetan Chiang and other close neighbours of the ancient Chinese gradually began 
to move into the central Chinese plain. In the year 308, 100 years before Alaric took Rome 
and before the first Barbarian Empire (the Kingdom of Toulouse) was established on the 
territory of the Roman Empire, the Hsiung-nu general Liu Yuan proclaimed himself emperor 
and three years later his successor, Liu Tsing, captured the capital of the Chin Empire and 
took prisoner the Son of Heaven. What is referred to here is the emergence of a dynasty of 
Hsiung-nu origin: the Former Chao (304-325). The Hsiungnu also founded the dynasties of 
the Northern Liang (397-439) and the Hsia (407- 431). Historians have called Liu Tsing the 
'Chinese Attila'. 16 

In the words of the ancient Chinese author, Chin-shu: 'The people is experiencing depriva- 
tion and is sad as a result; all have the same concern, all are waiting for peace and rest like 
dew and rain in a drought." 17 

The history of western Central Asia is also marked by a series of largescale nomadic 
movements from north and north-east to south. Of particular note is the movement of the 
Yiieh-chih in the second half of the second century B.C. which ultimately led to the for- 
mation of the powerful Kushan Empire, whose last days are described in this volume. The 
invasion of new ethnic groups led to the formation of the Chionite state, the Kidarites and 
subsequently the Empire of the Hephthalites, who moved into East Turkestan, Afghanistan 
and north-western India. Then came the invasion of the Turks, whose dominion was even 
more extensive than that of the Hephthalites. 

15 Maenchen-Helfen, 1973; Tikhvinskiy and Litvinsky, 1988. 

16 Kryukov et al., 1979. 

17 Materialy po istorii kochevykh narodov v Kitae, 1989, p. 158. 



ISBN 978-92-3-103211-0 Contents The Silk Route 

These successive waves of nomadic invasions are usually considered in ethno-political 
terms - this is understandable since such movements did indeed play a major, and at times 
decisive, role in the ethnic and political history of Central Asia. Large nomadic populations 
appeared in the vicinity of the oases and then within them. This had fundamental conse- 
quences: the nomadic sector of the economy grew in importance, as did the interdepen- 
dence - or even the intermeshing - of the economic contributions made by the nomadic 
and settled populations. Newly settled or former nomads appeared in settled rural and 
urban communities. Intensive inter-ethnic and linguistic processes developed alongside 
those of cultural synthesis and the mutual enrichment of cultures (see below). Members 
of nomadic clans were in overall charge of the state and of many domains; inter-ethnic 
marriages became common among the aristocracy. 

The Silk Route 

All these complex political and ethno-cultural processes developed with varying degrees of 
intensity, embracing part or all of the territory of Central Asia. Contacts between peoples 
grew with the development of commercial ties. In this connection, the Silk Route was 
of great importance. It is sometimes presented as something akin to a modern motorway 
linking different countries; but it was really a system of roads (and the principal direction 
in which they ran) rather than one specific road. There were also supplementary roads 
running close to the main road within each oasis and state. There were both land routes 
and sea routes. The entire network - running from China to the Mediterranean, over a vast 
expanse from the Yellow Sea to the central Mediterranean, from the southern Urals to the 
Indian Ocean - made up the Silk Route. 

In the second and first centuries B.C., the Silk Route had two branches through East 
Turkestan, running into western Central Asia and thereafter south to India and west through 
Iran and Mesopotamia to Antioch. For political reasons and also because of climatic 
changes, the network of towns altered, and the direction of the routes shifted as a result. 
From the fifth to the seventh century a.d., three roads ran through East Turkestan. The 
northern road led to Lake Issyk-kiil and then westward along the northern shores of the 
Caspian Sea, the Caucasus and the Black Sea to Asia Minor and Byzantium. The middle 
road crossed the Turfan depression and the northern rim of the Tarim basin in the direc- 
tion of the Ferghana valley, Samarkand, Bukhara and Merv and then ran through Iran to the 
eastern Mediterranean. The southern road ran from the area of Lop Nor through Khotan and 
Wakhan to Tokharistan, Bamiyan, northwestern India and thence by the sea route across 
the Indian Ocean to the Mediterranean. As one ancient Chinese author noted, 'There are 



ISBN 978-92-3-103211-0 Contents The Silk Route 

also roads running from each country which intersect in turn in the south and in the north. 
By following [these roads] it is possible to reach any point.' 18 

The path of many specific main and auxiliary roads has been established. At times they 
left unexpected traces: for example, stopping points for travellers have been discovered in 
the high mountains in Gilgit (Pakistan). There were also sanctuaries where they prayed. 
Here, too, the custom developed of carving a drawing or an inscription, a kind of Gdste- 
buch, in one of the cliffs. More than 10,000 rock drawings ( petroglyphs) have been dis- 
covered on the Karakorum road, as well as some 1,500 inscriptions in 17 languages and 24 
scripts, the largest number of inscriptions being in the Middle Iranian languages, mainly 
Sogdian. There are also Chinese, Indian, Hebrew and other inscriptions dating from the 
second, third and ninth centuries. 19 

As mentioned above, the Sogdians played the most important role in trade in Central 
Asia. By the fourth to the third century B.C., they had already begun to penetrate the eastern 
part of Central Asia. There were many populous Sogdian colonies at various points in 
East Turkestan, in Dunhuang, in China and in Mongolia, and large numbers of Sogdian 
merchants lived in Ch'ang-an, the capital of T'ang China. According to the texts discovered 
in East Turkestan and known as the ' Ancient Sogdian Letters' , the Sogdians of the diaspora 
did not lose their links with Samarkand. 20 

Trade on the Silk Route was often closely interwoven with politics. One episode involv- 
ing Turks, Sogdians, Iranians and Byzantines is typical. From the time of their arrival in 
western Central Asia, the Turks had shown an interest in the development of international 
trade and particularly in the colossal profits of the silk trade. This trade was conducted by 
the Sogdians through Iran but it was precisely in the Iranian sector of the route that the 
Central Asian (Sogdian) traders encountered the greatest difficulties. With the agreement 
of the Tiirks, the Sogdians themselves sent an embassy to Iran headed by the Sogdian, 
Maniakh (Menander fragment 18). The embassy proposed either that a through trade in 
silk to the Byzantine Empire should be permitted, or that the Persians themselves should 
purchase the silk from the Sogdians. These proposals were rejected. 

A second embassy was then dispatched with the same mission, this time by the Turk 
kaghan (king). Only a few members returned; the others perished in Iran, an indication 
that the shahanshah was preparing to initiate military action rather than engage in trade. 
In order to reach an agreement with Byzantium directed against Iran, the Tiirks dispatched 
yet another embassy (again led by Maniakh), which travelled along the northern shore of 

18 Herrmann, 1938; Raschke, 1978; Tikhvinskiy and Litvinsky, 1988. 

19 Dani, 1983a; 1983ft; Jettmar (ed.), 1989; Litvinsky, 1989. 

20 Henning, 1977; Pulleyblank, 1952; Harmatta, 1979; Grenet and Sims- Williams, 1987. 



ISBN 978-92-3-103211-0 Contents Cross-cultural influences 

the Caspian and through the Caucasus to Byzantium to pursue the question of the direct 
sale of silk to Byzantium. This embassy brought large quantities of silk and managed to 
conclude a Byzantine-Turk agreement directed against Iran. In 568 the delegation returned 
home, accompanied by a corresponding Byzantine embassy headed by Zemarkhos. Other 
embassies were then dispatched and the volume of trade increased considerably. 21 

The expression ' Silk Route' is perhaps a misnomer since much more than silk was 
traded along it. Lacquered ware, Chinese bronzes including (especially) mirrors, paper and 
much else from eastern Asia was sent to the West. It was in no sense a one-way road, as is 
often believed: a steady stream of goods was carried from the Mediterranean and Central 
Asia to the East and to China. The merchandise included cloth, silverware and coins, gold 
and gold artefacts, precious and semi-precious stones, glassware and livestock. There was 
also a considerable exchange of people. 22 An uninterrupted flow of Buddhist pilgrims from 
China and other regions travelled to and from India; in turn, Buddhist missionaries travelled 
from India to the most remote regions of Central Asia. A document found in Merv contains 
extracts from various Buddhist works compiled by a Buddhist missionary from Gilgit for 
his own use. Chinese merchants also transported their goods and a variety of books far to 
the west - traces of such activities have been found in the northern Caucasus. The Silk 
Route thus served for the movement not only of goods but also of ideas. 23 

Cross-cultural influences 

The discovery of written sources in East Turkestan provides clear evidence of the inten- 
sity of ethnic and cultural interaction. There are thousands of manuscripts in the Indian 
languages, Sanskrit and Prakrit, with the most varied content, both religious and secu- 
lar; and there have been rich finds of literary texts in Chinese and Tibetan. Many manu- 
scripts have been found in Iranian languages such as Middle Persian and Parthian, Sogdian, 
Khotanese Saka, Bactrian (Hephthalite) and also New Persian. Of great significance were 
documents in a previously unknown Indo-European language which has received the des- 
ignation Tokharian. Mention should also be made of the literary texts in Turkic, Syriac and 
other languages. 24 The area was an ethnic melting-pot, sometimes simmering quietly and 
at other times erupting: the reciprocal influence and intermingling of cultures was equally 

21 Pigulevskaya, 1951; Moravcsik, 1958. 

22 Schafer, 1963. 

23 Litvinsky, 1986. 

24 For further details, see Litvinsky, 1984. 



ISBN 978-92-3-103211-0 Contents Cross-cultural influences 

A dominant role was played by the region's various religions, which did not spread 
in isolation but brought with them a religious and cultural structure or sets of structures. 
Thus the spread of Buddhism, together with the ideas and principles of Buddhist architec- 
ture and iconography, led to the diffusion of Indian languages, scripts, philosophy, artistic 
works, astronomy, medicine and other sciences in addition to related moral and ethical 
principles. (To this should be added the influence of Hinduism.) The same remarks apply 
to Zoroastrianism, Christianity and Manichaeism. The latter incorporated many principles 
from other religions, particularly Buddhism. Taoism also spread from China. The nomadic 
peoples brought much from the world of their societies, including elements of religion and 
culture in a broader sense. Certain Hellenistic traditions were maintained. All these ele- 
ments intertwined with the equally varied local cultures and religions, producing clashes 
and interpenetration. Social structures also interacted in a parallel way. 

This multi-stage and multi-tiered cultural synthesis, the multiplicity of forms of polit- 
ical and social life, together with the emergence and development of feudal structures, 
characterize the contradictory and dynamic history of an age which produced remarkable 
values and an imposing civilization. 

Editors' Note 

Middle Asia is the territory belonging to the Commonwealth of Independent States (former 
Soviet Central Asian republics). 



ISBN 978-92-3-103211-0 





N. N. Chegini and A. V. Nikitin 



Struggles against the northern nomads 43 

The Sasanian administration 45 

Royal cities 45 

The reforms of Khusrau I 47 

The economy 48 

Sasanian coins 49 

The army, warfare and armaments 57 


Sasanian cities and fortifications 60 

Sasanian court architecture 63 

Sasanian religious architecture 66 

Sasanian art and crafts 67 

See Map 2. 




Part One 


(A/. N. Chegini) 

It is probable that Vologases IV died some time in a.d. 208/209, after which the throne 
of the Parthians was disputed between his sons, Ardavan V (Artabanus) and Vologases V. 1 
Ardavan ruled in central Iran and Vologases in Mesopotamia, striking coins at Seleukia. 2 
The conflict between the two brothers lasted until the end of Parthian rule. In Rome, Cara- 
calla succeeded his father Septimius Severus in 211 and the weakness of the Parthians 
resulted in a Roman incursion into Parthia, during which a great part of Media was pillaged 
and the Parthian tombs at Arbela were stripped. Although Ardavan succeeded in defeating 
the heir to Caracalla, Macrinus, the war against Rome and internal struggles strained the 
Parthian Empire to its limits. 3 What is now known, following Simonetta's work, is that 
Ardavan did not issue tetradrachms because he did not control Seleukia. 4 

After the invasions of Alexander the Great in the early fourth century B.C., the region 
of Fars, the homeland of the Persians, had become one of the vassal kingdoms of first the 
Seleucids and then the Parthians, ruled by several local princes. The kingdom of Persia 
issued coins almost continuously between 280 B.C. and a.d. 200, using the title prtrk' 
(Frataraka, i.e. governor) and later MLK' (king). By the beginning of the third century, 
conflict within the Parthian royal family and war with the Romans had weakened central 
authority. 5 

One prominent king of Persia during the last years of the weakened Parthia was Gochihr 
of the Bazrangi family, although his name does not appear on coins. 6 Sasan, after whom 
the dynasty is named, may have been the chief priest of the Adur Anahid temple in Istakhr. 

1 Simonetta, 1956, pp. 77-82. 

2 Ibid., p. 77. 

3 Ibid., pp. 77-8. 

4 Ibid., pp. 78-9. 

5 Frye, 1975, pp. 239-41; Sellwood, 1983, pp. 299-306; Frye, 1984, pp. 271-85. 

6 Al-Tabari, 1879-89, Vol. 1, p. 815; Herzfeld, 1924, pp. 35-6. 




Papak, his son or a descendant (as Sasan, although mentioned, does not appear in the 
family line of the Sasanians listed in Shapur's great inscription of the Ka c be of Zoroaster 
at Naqshi Rustam, near Persepolis), succeeded him and the family gained more authority 
by defeating the local governors and deposing Gochihr. Papak' s name appears on coins 
using the title MLK' ? According to the Arab historian al-Tabari, Ardavan V asked Papak 
to submit to his authority and to send his son Ardashir to the court, but he refused. When 
Papak died he was succeeded by his eldest son Shapur to whom Ardashir, the younger 
brother, did not give allegiance. When Shapur died in an accident in Persepolis, however, 
Ardashir became the head of the local dynasty. 8 

According to the Bishapur inscription of Shapur I, Ardashir proclaimed himself king in 
205. The series of coins showing him with a Parthian tiara probably commemorates this 
event. 9 Ardashir then campaigned in western Iran and conquered Susiana and Elymais in c. 
222. 10 Characene (Meshan), the vassal kingdom of the Parthians, was captured and a new 
governor appointed. At the famous battle of Hormizdagan (whose site is not known), 11 
which probably took place not later than 224, 12 Ardavan V was defeated and killed. 

After their crushing defeat, the remaining forces of the Arsacids (i.e. Parthians) fled to 
the mountains and resisted for a while. On the basis of evidence in the Mujmal al-tdwankh, 
Widengren suggests a second battle near Nihavend when Ardashir was marching towards 
the capital, Ctesiphon (Tespon). 13 According to al-Tabari, Ardashir advanced to Ecbatana 
(Hamadan) and then conquered Armenia and Adiabene (Mosul). In 226 he entered the 
capital and styled himself shahanshah (king of kings) and his official reign started. A 
commemorative bas-relief was ordered to be cut on the rock at Naqsh-i Rustam (Fig. 1) 
and coins showing him with a new crown were issued (Plate I, 3). 14 

It is now accepted that Ardashir I defeated Ardavan V several times, overthrew some 
of the minor local rulers who lived under the Parthians and replaced them with newly 
appointed governors from his own family. 15 If, however, al-Tabari's account 16 is correct, an 
eastern campaign must have taken place during the rule of Ardashir I, and Seistan (modern 
Sistan), Abarshahr (Nishapur), Merv, Balkh and Khwarizm were occupied. The overthrow 
of the Great Kushans, at least in the western part of their realm, is now considered the result 

7 Frye, 1984, p. 271; 1975, p. 239. 

8 Ghirshman, 1954, p. 290. 

9 Frye, 1983, p. 117. 

10 Al-Tabari, 1879-89, Vol. 1, p. 818; Hansman, 1978, p. 155. 

11 Rawlinson, 1876, p. 37. 

12 Bivar, 1969a, p. 50. 

13 Rawlinson, 1876, p. 37; Widengren, 1971, p. 743; Frye, 1975, p. 242. 

14 Henning, 1954, p. 44. 

15 Bivar, 1969a, p. 50. 

16 Ibid. 




FIG. 1 . Investiture relief of Ardashir I at Naqsh-i Rustam. (Courtesy of M. I. Mochiri.) 

of the rising power of the new dynasty in Iran. 17 According to al-Tabari, 18 the king of the 
Kushans (perhaps Vasudeva I) sent a mission of surrender. Whatever the circumstances, 
the kingdom of the Kushans was divided and the heartland of their empire in Bactria and 
the Kabul valley came under the control of the Sasanians. 19 The Sasanian rulers of the 
captured territory are known today as the Kushano-Sasanian governors, although the date 
when they began to issue coins is not known. 20 It seems that during this period the Sasanian 
kings regularly appointed governors of the principal provinces. 21 

The extent of Shapur I's empire in the east is known from the content of his inscription 
on the face of the Ka c be of Zoroaster. This inscription is written in three languages, Middle 
Persian, Parthian and Greek, and lists the provinces of the Sasanian Empire in c. 260. It 
shows that Shapur was already victorious against the Romans and in Transcaucasia, and 
under him Sasanian control in the east was also expanded. 22 The second part of the inscrip- 
tion, which is a description of the empire, gives Shapur's possessions as Merv, Herat and 
all Abarshahr, Kerman, Seistan, Turan, Makran, Paradan, Hind (Sind) and Kushanshahr 
as far as pshkbwr (Peshawar) and up to the borders of Kash, Sughd and Chach (Kashgar, 

17 Frye, 1975, p. 242. 

18 Bivar, 1969a, p. 50. 

19 Ibid. 

20 Bivar, 1979, pp. 317-32. 

21 Ibid., pp. 320, 323-4. 

22 Henning, 1937-39; 1954, pp. 




ISBN 978-92-3-10321 1-0 Contents Struggles against the northern nomads 

Sogdiana and Tashkent). This passage also lists all the provinces situated in the east of the 
empire. 23 In mentioning the Kushans, Shapur indicates the extent of his control to the east 
and north-east. It should be pointed out that the land of Khwarizm, although not appear- 
ing in the list of provinces, had already been captured by the Sasanians during the rule 
of Ardashir I. Al-Tabari mentions a campaign in which Ardashir conquered Khwarizm as 
well as Gurgan, Merv and so on. According to the Chronicle ofArbela (whose authenticity 
is open to doubt), the final assault on Khwarizm took place in 239/240 during Ardashir's 

The appearance of Kushanshah (king of the Kushans) as a Sasanian title shows that 
the Great Kushan kings had been defeated. Shapur I's success evidently ended the rule of 
the Great Kushans and split their kingdom into two parts, the northern and the southern. 
Branches of the Kushans ruled in the southern part, east of the River Indus, where they 
are known as Murundas. 24 The northern part, or core, of the Kushan territory became a 
province of the Sasanian Empire. 

Struggles against the northern nomads 

Shapur II (309-379) was forced to wage war for ten years against invaders whom Ammi- 
anus Marcellinus (XVII, 5) refers to as the Chionites. Shapur was clearly successful in 
his operation and managed to impose his authority on the invaders and stabilize his east- 
ern frontiers. 25 The victorious return of Shapur must have taken place some time before 
360; it was apparently at this time that the city of Abarshahr was founded and used as his 
headquarters. 26 His success in containing the Chionites resulted in the conclusion of an 
alliance under the terms of which the Chionites would help Shapur in his war against the 
Romans. In 360, when he laid siege to the fortress of Amida (the modern Diyarbekir in 
eastern Turkey), the Chionites with their king Grumbates supported him, according to the 
eyewitness account of Ammianus Marcellinus (XVIII, 7. 1-2). 27 

A few decades later, it appears that Kushanshahr was no longer under the control of the 
Sasanians and was subject to new invaders. 28 This new power, known to us as the Kidarites 
(after their leader Kidara, probably himself a Chionite chief), had appeared on the eastern 
frontiers by the end of the fourth century. Coins of Kidara, together with those of Shapur II, 

23 Herzfeld, 1947, p. 182; Henning, 1947-48, p. 54. 

24 Bivar, 1969a, p. 51. 

Ti T1 • 1 ro A 

25 Ibid., pp. 53-4. 

26 Ibid., p. 53. 

27 Ibid., p. 54; 1979, p. 327. 

28 Bivar, 1969a, p. 54. 



ISBN 978-92-3-10321 1-0 Contents Struggles against the northern nomads 

Ardashir II and Shapur III, have been found in the treasure of Tepe Maranjan near Kabul 29 
and in the archaeological site of Butkara, Swat (in Pakistan). The Kidarites (who dominated 
Tokharistan and Gandhara) adopted the Sasanian title of Kushanshah, which indicates that 
they were the chief heirs of the Sasanian Kushanshah?, and their administration. 30 We know 
that the new wave of invaders from Iran came at the time of Bahram V shortly before 
440. It is reasonable to suppose that this new disturbance was caused by the arrival of the 
Hephthalites and that early in the fifth century they drove the Kidarites south from Bactria 
to Panjab, where the name Kidara appears on many gold coins. 31 

It is clear that from early in the fifth century, the Hephthalites had become the main 
power in the east: it was to them that the Sasanian prince Peroz appealed for assistance in 
defeating his brother Hormizd III in 457. Although Peroz (459-484) succeeded in recov- 
ering his throne, he was later defeated and captured by his former allies. According to 
al-Tabari, the name of the Hephthalite king was Akhshunvar (a Sogdian title, khsundar 
meaning 'king'); or Khushnavaz according to the poet Firdausi. Peroz was freed in return 
for leaving his son Kavad as hostage; and when Kavad was ransomed, Peroz returned and 
attacked the Hephthalites. This resulted in his defeat and death, and the loss of his army. 32 
After this defeat the Sasanians had to pay an annual tribute to the Hephthalites and some 
parts of the eastern region fell into the hands of the enemy. Kavad even asked the Romans 
to lend him money to pay the tribute. 33 In 498 or 499, however, it was through Hephthalite 
support that Kavad I regained his throne. 34 

During the rule of Khusrau I Anushirvan (531-579) the Turks arrived on the Jaxartes 
steppes from Mongolia. In order to crush the Hephthalites, Khusrau allied himself with the 
Turk kaghan known in Arabic and Byzantine sources as Sinjibu or Silzibul. A fierce battle 
took place, the result of which was the defeat and dispersal of the Hephthalites and the 
division of their land. The southern part was taken by the Sasanians and the north by the 
Tiirks. 35 At the same time Khusrau I rebuilt the lines of fortification on the Gurgan plain 
of eastern Mazandaran. One such fortification was Sadd-i Iskandar (Alexander's barrier'), 
or Sadd-i Anushirvan; a second was the wall of Tammisha, running from the mountains to 

29 Bivar, 1979, p. 331; Curiel, 1953, pp. 107-9. 

30 Gobi, 1976, p. 340; Bivar, 1979, p. 331. 

31 Bivar, 1969a, p. 55. 

32 Ibid. 

33 Christensen, 1944, pp. 316-17. 

34 Bivar, 1969a, p. 56. 

35 Ibid.; 1975; 1983, p. 215. 



ISBN 978-92-3-103211-0 Contents Royal cities 

the seashore and closing the eastern approach to Mazandaran. Khusrau is also supposed to 
have rebuilt the wall and defences of Darband in the Caucasus. 36 

The Sasanian administration 

At the head of the Sasanian state stood the king. In official inscriptions the Sasanian kings 
called themselves 'Mazda - worshipping majesty, of the race of the gods'. According to 
Ammianus Marcellinus (XXIII, 6.5), the Sasanian king considered himself 'brother of the 
sun and moon'. On reliefs, 'in the language of transparent symbols, the King of Kings is 
shown as the earthly incarnation of the supreme deity'. 37 

During the Early Sasanian period the administration of the provinces and districts did 
not differ greatly from that under the Parthians. It was during this period that the royal 
cities, almost equivalent to semi-independent kingdoms, were built (see below). 38 In the 
early inscriptions we find mentions of shahrs (vassal kingdoms) such as Merv, Kerman, 
Sakastan, Adiabene, Iberia, Makran, Mesene, Kushanshahr and Armenia, which had sub- 
mitted to Sasanian rule. In many cases the rulers of these kingdoms were the sons of the 
monarch himself. 

In the Early Sasanian period, Shapur I was the ruler of the kingdoms listed, all of which 
had to pay tribute and submit in varying degrees. 39 It was in the later part of the Sasanian 
period that a greater centralization took place: in theory the empire was divided into four 
parts, each governed by an official appointed by the king, with both military and civil 
powers. The title of the commander was spdhbad. 40 

Royal cities 

The Sasanian royal cities (under the administration of a shahrab) were the headquarters 
of the military garrison, centres of newly formed administrative districts and residences of 
the state officials. 41 Ardashir I himself founded many cities, one of which was Ardashir- 
Khvarreh ('Glory - or fortune - of Ardashir'). From a military outpost, it grew to become 
an administrative district with Gur as its centre. It was laid out on a circular urban plan. 
In the words of al-Tabari, 'Shapur I, like his father, founded or renamed cities and we can 

36 Frye, 1977, pp. 7-15; Bivar and Fehrevari, 1966, pp. 40-1, PI. 11 (a-b) and Fig. 1 (map of the 
region) ;Kiani, 1982a, pp. 73-9; 1982ft; Bivar, 1983, p. 215. 

37 Frye, 1983, p. 160. 

38 Lukonin, 1983, p. 735. 

39 Ibid., pp. 731-2. 

40 Ibid., pp. 723-5. 


Ibid., pp. 120-1, 162, 751, 1056. 



ISBN 978-92-3-103211-0 


Royal cities 

see an example of both in his inscription - Gundeshapur and Peroz-Shapur - while other 
towns mentioned by Arabic or Persian authors may be attributed to either Shapur I or II.' 42 
According to Christensen, 'Other cities were Shad-Shapur, "Joyful is Shapur", or c Ubulla 
in southern Iraq, Shapur-Khvast near Khurramabad, Vuzurg-Shapur or c Ukbara in Iraq, as 
well as others, but none in the eastern part of the empire. These cities, like Darabgird and 
Gur in Fars, were surrounded by walls and were presumably well fortified, a feature of 
Sasanian city planning.' 43 The most famous city founded by Shapur I was Bishapur, with 
a Greek plan. It was probably built in a.d. 262, six years after his triumph over the Roman 
emperor Valerian. 

The administrative capital of the Sasanians was Ctesiphon. It consisted of a group of 
towns known as the madd'in (meaning 'the cities' in Arabic), two of which were Veh- 
Ardashir and Veh-Antiokh-Khusrau; the district in which they were situated was called 
(at least during the sixth century) Shad-Kavad. Taq-i Kasra (Fig. 2), a building dating 
probably from the Early Sasanian epoch and extended or embellished during the rule of 
Khusrau I, was situated in the city. Ctesiphon was not only the seat of most Sasanian 
kings but also the most important of the Sasanian capitals in economic and strategic terms. 
Besides cities such as Ardashir-Khvarreh, Bishapur, Gundeshapur, Susa, Dastagird (held 

FIG. 2. Taq-i Kasra at Ctesiphon. (Photo: © Barbara Grunewald.) 

42 Al-Tabari, 1879-89, Vol. 1, pp. 961-2. 

43 Christensen, 1944, p. 361. 



ISBN 978-92-3-10321 1-0 Contents The reforms of Khusrau I 

as a capital during the reign of Khusrau II and located east of Ctesiphon) and Ecbatana 
(a summer capital), the city of Istakhr in Fars also served as an administrative, religious 
and economic centre. It was the ideological heart of the empire, since the temple of the 
dynasty's fire - the coronation place of many Sasanian rulers - was situated there. 

The reforms of Khusrau I 

Khusrau's success in overcoming the religious movement of the Mazdakites (see Chapter 
17) and managing to put the country's life in order gave him a great opportunity to start 
his reforms. One of these was the new policy on land taxation and the poll tax. According 
to al-Tabari, the change in the land tax had already begun during the reign of Kavad I. 44 
The Arab historian reports that the farmers had had no right to harvest crops or gather 
fruit from their garden before the arrival of the tax collector; the long wait meant that their 
produce was frequently wasted. To avoid this, Khusrau introduced a new fiscal system. 
First, he ordered the lands to be measured. Next he fixed the amount of tax to be levied for 
each grlv (Arabic jarlb, one-tenth of a hectare) according to what was cultivated there; for 
example, 1 grlv of wheat or barley = 1 drachm, and 1 grlv of vineyard = 8 drachms. Under 
the new regulations, all persons between the ages of 20 and 50, except nobles, soldiers and 
priests, were compelled to pay the poll tax, whose amount ranged from 12 drachms (Arabic 
dirhams) to 4 drachms, according to wealth. 

The taxes were collected according to the administrative sub-divisions of the country 
from village up to province, with the officer in charge of a province being responsible 
for overall supervision and the tax in each city being paid to the judge of that city. Al- 
Tabari reports that Khusrau ordered the list of the new tax rates to be kept in the royal 
treasury. With the implementation of these new measures and the appearance of organized 
tax collectors, Khusrau was able to maintain a regular income for the government. Super- 
vision of the payment of taxes into the royal treasury was undertaken by the hdmdrkar 
(accountant), who was also responsible for issuing documents on the right of ownership 
and possession. 45 The great tax reform of Khusrau I marked a turning point in the Sasanian 
state administration. For the first time, the power of the landed nobility was restricted and 
all the taxes were in the hands of the king. 46 

44 Al-Tabari, 1879-89, Vol. 1, pp. 961-2. 

45 Lukonin, 1983, p. 726; Dennet, 1950, p. 15. 

46 Al-Tabari, 1879-89, Vol. 1, p. 963; Christensen, 1944, p. 362; Lukonin, 1983, p. 746. 



ISBN 978-92-3-103211-0 Contents The economy 

The economy 

Since the vast majority of the population were peasants, the country's economy was based 
on land and agriculture. The archaeological survey of Khuzistan and the area north of 
Baghdad shows the great Sasanian interest in irrigation and cultivation. One of the great 
irrigation systems was the Nahravan canal, which supplied the water for a vast area of 
cultivation. The remains of Sasanian canals and dams can still be seen in various parts of 
Iran. These activities increased during the rule of Khusrau I, under whom large areas of 
land were brought under cultivation. 47 

Thus the national economy continued to be based on agriculture rather than trade. In 
commerce, Sasanian coinage of silver and copper, more rarely of gold, circulated over a 
wide area and the bill of exchange appeared. 48 More money was in circulation in the towns, 
as shown by the great number of silver drachms found in Iran and neighbouring countries. 
In the rural districts, however, the wages of the peasants, soldiers and officials, and even 
some of the taxes, were paid in kind. The levying of dues and taxes in kind enabled the 
government to build up large stocks of essential goods that could be called upon in time of 
famine. 49 

It is probable that silk was already being imported into Iran from China in Achaemenid 
times. 50 In the Sasanian era, two routes were used, one overland (still called the Silk Route) 
and the other the sea route around the coasts of South-East Asia, although this was less 
popular than the overland route. 51 Silk was woven mainly in Syro-Phoenician and Chinese 
workshops; besides the woven silk from China, large quantities of raw silk yarn were also 
imported for weaving to purely Sasanian designs, creating a rival industry. The workshops 
of Susa, Gundeshapur and Shushtar were later famous for their products. 52 

Luxury ceramics, glassware, textiles, amber and papyrus were imported and there was a 
transit trade in spices from China and Arabia. 53 However, Iran's position was as a middle- 
man that benefited from the value of the traded items. The excavated finds from Begram, 
which can be ascribed to the Early Sasanian period, indicate commodities in transit such 
as decorated glassware and glass beads, ivories and manufactured metal. Although the 

47 Frye, 1983, pp. 160-1; on Khuzistan, see Adams, 1961; on Iraq, see Adams, 1965. 

48 Ghirshman, 1978, pp. 341-2. 

49 Ibid. 

50 Bivar, 1970, p. 1. 

51 Ibid. 

52 Ghirshman, 1954, p. 342; Bivar, 1970, p. 2. 

53 Ghirshman, 1954, p. 342. 



ISBN 978-92-3-103211-0 Contents Sasanian coins 

Sasanian coin finds from China show the use of Iranian silver, this is not enough to prove 
that Sasanian imports from China were substantially financed by a mass of silver 



Sasanian coins 

All the Sasanian rulers struck coins, and these are an invaluable source of historical, 
cultural and economic information. A constant denomination and weight standard were 
adopted, and the coins bore the ruler's effigy on the obverse and a fire-altar on the reverse. 
On the obverse the king's portrait faces right, in contrast to the practice under the Arsacids 
- except on commemorative issues, frontal portraiture is rare (Plate V, 29). The name of the 
king and his titles are inscribed close to the edge. Each ruler has his own personal crown, 
which is a reliable guide to the whole range of Sasanian art and its chronology. Only one 
queen's portrait, that of Boran, appears on the coins (Plate VI, 35). 

On the reverse the fire-altar with flames always appears, with three principal variations: 
by itself (Plate I, 1,2, 3), with two flanking figures, or with a bust in the flames (Plate 
III, 14, 15). The significance of the two attendant figures in the second type has not yet 
been clarified. At the beginning of the issues the figures carry long rods, and later barsom- 
bundles in their hands, facing towards or away from the altar. In the time of Khusrau I 
they appear frontally (Plate IV, 23), and from the time of Bahram II (275-293) onwards at 
least one of the attendant figures, judging from the crown, represents the ruler (Plate II, 7, 
8). Special reverse designs allude to investitures (Mithra or Anahita); special issues under 
Khusrau II (590-628) show a bust in a nimbus of flames (Plate V, 29), or the king standing. 

The Sasanians adopted the traditional silver drachm of Attic weight, the most common 
currency of the Parthian period. The weight (almost 4 g) and the fineness of the metal used 
were, with few exceptions, well maintained. Besides drachms there were half-drachms, 
obols and half-obols, and tetradrachms of a poor silver alloy (billon). The striking of gold 
was also revived, but only for prestige and display issues. Some rulers did not strike gold 
coins and after Khusrau II their issue ceases. 55 The formulation of Sasanian coin inscrip- 
tions was determined by the political and religious motives of the dynasty. They are written 
in Sasanian Pahlavi with the use of ideograms. On the obverse the royal name and titles 
appear, and on the reverse the name of the royal fire, with later the place of minting and 
regnal year. 56 

54 Bivar, 1970, pp. 2-4. 

55 Gobi, 1983, pp. 322-8. 

56 Ibid. 



ISBN 978-92-3-103211-0 


Sasanian coins 

Pigs. i.3,J. AWaJiir I 

FlG- 4. Stapur I. Rg. i HumnzJ. I FtG. i Gahrun I 

PLATE I. (Courtesy of M. I. Mochiri.) 



ISBN 978-92-3-103211-0 


Sasanian coins 

FHH, ?.«. Balirtjnill. 

! '•? >■ :-i|i 

Fit- 10. Ni«*h 

Pit II- Htimwd It 

(golJ dirur)- 

PLATE II. (Courtesy of M. I. Mochiri.) 



ISBN 978-92-3-103211-0 


Sasanian coins 

F: 17. Ardaihir II 

Fl> 14. Stiapur III FlS. IS. Baliraii IV 

Ft;, 14 Yudginl I Peg 17. Bahram V Tig. IS. Yxz^irJ CL 

PLATE III. (Courtesy of M. I. Mochiri.) 



ISBN 978-92-3-103211-0 


Sasanian coins 

PIC 19. FcraiL 

Fig 20. Vilasli 

Fig 21 Kami I 

Pit 12 Jimup. PlG. £J. fchuinii I. FtC, 34. HcnrezJ IV 

PLATE IV. (Courtesy of M. I. Mochiri.) 



ISBN 978-92-3-103211-0 


Sasanian coins 

Pia.29 EihnmM 

FIG. 26. V»tjh*n 

PlG* 37. ] jormiid V 

PLATE V. (Courtesy of M. I. Mochiri.) 



ISBN 978-92-3-103211-0 


Sasanian coins 

PlC. J1. Kavad I] 

FLOS, K, J J. ArdatWrHL 

F]ii >4. Kliutniii EIL Fts, JS. Quctn Bonn FlG J*. Quren Aunnigidtikht 

PLATE VI. (Courtesy of M. I. Mochiri.) 



ISBN 978-92-3-103211-0 


Sasanian coins 



PI'S- 37. HgnncdVt 

hi - .*S Khusrau X\ 

FlGi J*. 4<i, 41 Vadpid m 
PLATE VII. (Courtesy of M. I. Mochiri.) 



ISBN 978-92-3-10321 1-0 Contents The army, warfare and armaments 

One of the chief characteristics of the Late Sasanian coinage is represented by the mint 
monograms that appear on the right side of the reverse and give the names of mint cities in 
abbreviated forms. These forms stand for the full Sasanian names of those cities authorized 
to possess mints. About 200 of these mint signs are known to us, 57 such as: 

'W for Hormizd Ardashir (present-day Ahvaz) 

'PR for Abarshahr (Nishapur) 

'RT for Ardashir-Khvarreh (Firuzabad) 

'HM for Hamadan 

'YL'N for Eran-Khvarreh-Shapur ( Susa) 

BYS for Bishapur in Fars. 

The army, warfare and armaments 

Khusrau Fs second most important reform was the reorganization of the army which, 
together with the implementation of the new taxation system, gave him a secure founda- 
tion from which to safeguard the empire. Previously, all nobles, great and small, had been 
obliged to equip themselves and their followers and serve in the army without pay, but 
Khusrau issued equipment to the poorer nobles and paid a salary for their services. Conse- 
quently, the power of the great nobles - who frequently had their own private armies - was 


reduced. A permanent army of cavalry known as aswaran (Arabic, asawira), designat- 
ing a heavily armed and disciplined force, existed. 59 According to Ammianus Marcellinus 
(XXIII, 6.83): 'They rely especially on the valour of their cavalry, in which all the nobles 
and men of rank undergo hard service.' 

The weaker part of the army was the infantry, consisting of peasants subject to military 
service. Ammianus Marcellinus reports (XXIII, 6.83; XXIV, 6.8), 'The infantry are armed 
like the murmillones [gladiators], and they obey orders like so many horse-boys.' Accord- 
ing to Procopius, the Persian infantry were used to destroy town walls after a victory. There 
were also auxiliary troops from the various nations allied to the central government such 
as the Armenians, the Hephthalites and the Dailamites. 60 Al-Tabari tells how a group of 
Dailamites under the command of Vahriz were sent to capture the land of Yemen. 61 The 
command structure of the army was also changed under Khusrau I. Previously the entire 

57 Bivar, 1963, pp. 169 et seq.; Tyler-Smith, 1983, pp. 240-7 (review of Mochiri, 1977). 

58 Frye, 1983, p. 155. 

59 Al-Tabari, 1879-89, Vol. 1, p. 2562; Vol. 2, p. 1604. 

60 Inostrantsev, 1926, pp. 23^1. 

61 Al-Tabari, 1879-89, Vol. 1, pp. 948-88. 



ISBN 978-92-3-10321 1-0 Contents The army, warfare and armaments 

army had been under the command of an officer known as the spdhbad. Now, four com- 
manders were appointed, each in charge of the troops of one-quarter of the country. 62 Each 
of these newly created commanders had a deputy called a marzban. The soldiers were 
inspected every year in order to prevent them from escaping their duty, and to maintain 
their equipment. Conditions of service were arduous and all soldiers had to study and be 
familiar with a range of military instructions, information on which can be found in the 
Pahlavi book, the Denkard. 63 

A fragment of a military treatise found in the c Uyun al-akhbdr (Ibn Qutaiba Dinawari) 
confirms the existence of a military book during the reign of Khusrau I, who himself may 
have written such a treatise. Al-Tabari 64 records the equipment that members of cavalry 
units were required to carry at muster parades of Khusrau I: mail, breastplate, helmet, leg 
guards, arm guards, horse armour, lance, buckler, sword, mace, battle axe, quiver of thirty 
arrows, bow case with two bows, and two spare bow strings. According to Frye: 

The reform of the army. . . was changed from the previous practice of the great feudal lords 
providing their own equipment and bringing their followers and retainers into the field to 
another system with a new force of dihqdns or 'knights' paid and equipped by the central 
government. . . Also, it should be remarked that the army reorganization under Chosroes 
[Khusrau I] was concentrated on organization and on training, rather than any new weapons 
or technical advances, and as previously the heavily armed cavalry remained the dominant 
force with archers less important. The masses, as usual, were still camp followers and little 
more than a rabble looking for booty, but a new nobility of service was created which became 
more influential than the landed nobility. Since payment in specie or even in kind did not 
suffice to recompense the 'knights', villages were granted to them in fief, and a large class of 
small landowners came into existence. . . Walls and forts also were built on the frontiers. 65 

The Sasanians expended great effort in fighting Rome, Byzantium and the eastern nomads 
who invaded the Iranian frontiers. They clearly had a strong and efficient military force. 66 
There were changes in the conduct of warfare over time, however, one of which was the 
development of the bow as a primary weapon with the arrival of the Huns during the mid- 
fourth century. 67 

The relief frieze of Ardashir I at Firuzabad is a representation of cavalry warfare. It is 
known that the king personally took part in the battle. 68 Although the Ardashir relief does 

62 Ibid., p. 894; Christensen, 1944, p. 365. 

63 West, 1982, pp. 86-90. 

64 Al-Tabari, 1879-89, Vol. 1, p. 964; Bal c ami, 1874, p. 1048; Christensen, 1944, pp. 362-3; Bivar, 1972, 
pp. 276-91. 

65 Frye, 1984, p. 326; on the military reforms, see Bivar, 1972, pp. 273-91. 

66 Bivar, 1972, pp. 273-91. 

67 Ibid., p. 281. 

68 Ibid., p. 275. 



ISBN 978-92-3-10321 1-0 Contents CUSTOMS, ARTS AND CRAFTS 

not depict the mace and battle axe, there is evidence of their use during the Late Parthian 
and Early Sasanian periods. The cavalry do not appear to have used shields in the Early 
Sasanian period. 69 

Part Two 

(A V. Nikitin) 

One of the best descriptions of Iranian customs and lifestyle in the Sasanian period is 
that given by Ammianus Marcellinus (XXIII, 6.75-80): 

Among these many men of differing tongues there are varieties of persons, as well as of 
places. But, to describe their bodily characteristics and their customs in general, they are 
almost all slender, somewhat dark, or of a leaden pallor, with eyes grim as goats', eyebrows 
joined and curved in the form of a half-circle, not uncomely beards, and long, shaggy hair. All 
of them without exception, even at banquets and on festal days, appear girt with swords; an 
old Greek custom which, according to the trustworthy testimony of Thucydides, the Atheni- 
ans were the first to abandon. Most of them are extravagantly given to venery, and are hardly 
contented with a multitude of concubines; they are free from immoral relations with boys. 
Each man according to his means contracts many or few marriages, whence their affection, 
divided as it is among various objects, grows cold. They avoid as they would the plague 
splendid and luxurious banquets, and especially, excessive drinking. Except for the kings' 
tables, they have no fixed hours for meal-times, but every man's belly is, as it were, his sun- 
dial; when this gives the call, they eat whatever is at hand, and no one, after he is satisfied, 
loads himself with superfluous food. They are immensely moderate and cautious, so much 
so that they sometimes march through an enemy's gardens and vineyards without coveting 
or touching anything, through fear of poison or magic arts. Besides this, one seldom sees a 
Persian stop to pass water or step aside in response to a call of nature; so scrupulously do 
they avoid these and other unseemly actions. On the other hand, they are so free and easy, 
and stroll about with such a loose and unsteady gait, that one might think them effeminate; 
but, in fact, they are most gallant warriors, although rather crafty than courageous, and to be 
feared only at a long range. They are given to empty words, and talk madly and extravagantly. 
They are boastful, harsh and offensive, threatening in adversity and prosperity alike, crafty, 
haughty, cruel, claiming the power of life and death over slaves and commons. They flay men 


Ibid., p. 276. 



ISBN 978-92-3-10321 1-0 Contents Sasanian cities and fortifications 

alive, either bit by bit or all at once, and no servant who waits upon them, or stands at table, 
is allowed to open his mouth, either to speak or to spit; to such a degree, after the skins are 
spread, are the mouths of all fettered. 

This picture is supplemented by the surviving Sasanian works of art, most of which depict 
scenes from the lives of kings or noblemen. Life at the royal court was governed by a 
strict code of protocol: the Byzantine system of court etiquette borrowed much from the 
court of the Iranian shahanshah (king of kings). 70 Sasanian inscriptions enumerating the 
members of the royal family and courtiers, and the positioning of the king and nobles on 
reliefs, are in a standard order. Arab and Byzantine sources provide descriptions of the 
ceremony for the reception of ambassadors to the court of the shahanshah?,. One of the 
traditional pastimes of the king and nobles was the hunt, for which special preserves or 
game parks were built. Ammianus Marcellinus (XXIV, 5.2) describes one such park that 
Roman soldiers saw during the emperor Julian's campaign in Mesopotamia in the year 
363. Scenes of the royal hunt were the most common theme used to ornament Sasanian 
silverware (Fig. 3). 

Sasanian cities and fortifications 

So far, there has been adequate archaeological investigation of only a small number of 
the cities of Sasanian Iran: consequently, there are few examples from which to assess the 
system of urban design. The layout of many of the ancient Parthian cities appears to have 
survived unchanged into Sasanian times. Late tradition ascribes the founding of a large 
number of cities to Ardashir I (226-241). The new capital of Ardashir-Khvarreh (modern 
Firuzabad) built by Ardashir in Fars was of archaic layout. Circular in ground plan, it was 
surrounded by a wall with four gateways placed at the points of the compass. The various 
districts were delineated by main streets radiating from the centre and dividing the city into 
20 sectors. 71 

Bishapur, built by Shapur I (241-271) in the seventh decade of the third century, had 
a regular ground plan in keeping with the rules for city design elaborated by Hippodamus 
of Miletus. Two main streets intersecting at right angles divided the city into four main 
districts, which were also of a regular layout. The architecture of Bishapur reveals clear 
Roman and Iranian influences and the Greek and Syrian masters and artists whom Shapur 
invited also had a hand in its construction. 72 By contrast, the port city of Siraf on the coast 
of the Persian Gulf grew up without a master plan. Its districts, which occupied an area of 

70 Lukonin, 1983, p. 710. 

71 Huff, 1973, p. 193. 

72 Ghirshman, 1956, p. 194. 



ISBN 978-92-3-103211-0 


Sasanian cities and fortifications 

FIG. 3. Sasanian silver dish showing a royal hunt. Photo: © The Trustees of the British Museum. 

around 1 sq. km, lay against the fortress that protected the port and were surrounded by a 
wall. 73 The irregular layout of some contemporary Iranian cities would appear to date back 
to Sasanian times. 

Most Sasanian cities were fortified. A system of fortresses and forts protected the bor- 
ders of Iran, the approaches to the cities and the caravan roads. Their walls were built of 
stone blocks, cemented cobbles or sunbaked brick. At the corners, and at regular intervals 
along the walls, there were round towers with narrow, vertical arrow-slots. The tower of 
the Sasanian fort at Turengtepe in northern Iran had eight arrow-slots on the lower tier, 
widening from their mouths (Fig. 4). 74 Studies have been undertaken of the fort in Siraf, 

73 Whitehouse and Williamson, 1973, p. 35. 

74 Deshayes, 1973, p. 144. 



ISBN 978-92-3-103211-0 


Sasanian cities and fortifications 

FIG. 4. Tureng-tepe. Sasanian fort. (Photo: © Mission Archeologique Francaise en Iran.) 

FIG. 5. Takht-I Sulaiman. General view. Photo: © Deutsches Archaologisches Institut, Eurasien- 

probably built under Shapur II (309-379) to protect the port against Arab attacks: its stone 
walls with round towers form a square, each side measuring 50 m; and the entrance in the 
middle of the south wall is defended by a barbican. 75 


Whitehouse and Williamson, 1973. 



ISBN 978-92-3-103211-0 Contents Sasanian court architecture 

Some idea of fifth- and sixth-century fortifications is given by the walls and towers of 
Takht-i Sulaiman (formerly Shiz) in Iranian Azerbaijan (Fig. 5). Built of dressed stone 
embedded in cement, and crowned with battlements, the walls were about 4 m thick and 
reached a height of 12 or 15 m. 76 The remains of fortifications built in the Parthian period 
and rebuilt in the fifth and sixth centuries still stand near Gurgan: they once defended Iran's 
northern border against nomadic raids. A system of fortresses set every 3-6 km, and linked 
by an unbroken wall, stretched from the Caspian shore more than 180 km eastwards. 77 

The acme of Sasanian military construction is represented by the fortifications of Dar- 
band, which stood across the road along the west coast of the Caspian; their construction 
began under Yazdgird II (438-457). The defences include the city's northern and southern 
walls, the citadel and a wall strengthened by stone forts that stretched 40 km to the Cauca- 
sus mountains. The fortifications were originally built of sunbaked brick and later of raw 
stone, but in the mid-sixth century, under Khusrau I (531-579), new walls were built of 
large stone slabs on the old foundations. To this day, the walls of Darband stand 20 m high 
in places. 78 

According to tradition, Shapur I used Roman prisoners of war to build dams and bridges 
in Mesopotamia and Khuzistan. It appears to have been during his reign that the irrigation 
works were constructed on the banks of the Karun river near Shushtar; the most famous is 
the Valerian dam and bridge. Faced with stone slabs, this enabled the water level in one of 
the Karun's tributaries to be raised by 2 or 3 m. The length of the stone bridge, which is 
reminiscent of similar Roman structures, was over 500 m. Another bridge more than 400 
m long has been preserved in Dezful, not far from Susa. There were several bridges across 
ravines and rivers on the ancient road that linked Ecbatana (modern Hamadan) and Susa. 

Sasanian court architecture 

Sasanian court architecture differed considerably from that of the Hellenistic and Parthian 
periods. From the outset it made use of a number of new principles that were then retained 
in later times. Some of them probably originate in the architectural tradition of Fars, which 
is as yet little known. In a number of cases a conscious imitation of Achaemenid models 
and Roman or Parthian influence are to be seen. The earliest monument of Sasanian court 
architecture is the palace of Ardashir I near Firuzabad. 79 Its design observes the principle 
of combining the apadana (official palace) and the harem (private residence). Measuring 

76 Naumann, 1977, p. 34. 

77 Kiani, 1982&, p. 12. 

78 Kudryavtsev, 1982, p. 65. 

79 Reuter, 1938, p. 534. 



ISBN 978-92-3-103211-0 


Sasanian court architecture 

55 x 104 m, the palace is laid out along the north-south axis. The northern part contains the 
throne room and the large rooms alongside it, roofed with low cupolas on squinches. The 
anteroom to the throne room is of the open aiwdn (hall) type. The southern part contains 
the inner courtyard onto which the private rooms look out. The plaster decoration of the 
niches imitates the stone ornamentation on the portals of Achaemenid palaces. The walls, 
which are up to 4 m thick, are made of stone cemented with lime mortar. 

Shapur Fs palace at Bishapur reflects Graeco-Roman architectural influence. A cruci- 
form throne room is contained within a square of outer walls 50 x 50 m, with a narrow 
gallery around the perimeter. The plaster decor - the meander and the plant designs - imi- 
tates the typical motifs of the art of Imperial Rome. The mosaics adorning the floor of the 
Grand Aiwdn are in the Syro- Roman style - the dancing girls, musicians and theatrical 
masks depicted scenes from the Dionysiac cycle. 80 

The combination of aiwdn and domed premises was also used in later palaces. The 
same design was used for the palace at Sarvestan near Firuzabad (Figs. 6 and 7), which 
tradition ascribes to Bahram V (420-438). It has a more sophisticated roofing structure, 
however: the vaults of the two symmetrically located side rooms rest not on walls, but on a 
system of arches standing on square columns (Fig. 8). 81 The next development in this tradi- 
tional design is represented by the palace of Aiwan-i Karkha near Susa, built under Kavad 
I (488-531). Its central hall, roofed with a cupola, is an open pavilion, the walls being 

flr „ : 

FIG. 6. Palace of Sarvestan. General view. Photo: © Deutsches Archaologisches Institut, Eurasien- 

80 Ghirshman, 1956, p. 193. 

81 Shepherd, 1983, p. 1065. 



ISBN 978-92-3-103211-0 


Sasanian court architecture 

FIG. 7. Palace of Sarvestan. General view. Photo: © Deutsches Archaologisches Institut, Eurasien- 

FIG. 8. Roofing structure at the palace of Sarvestan. Photo: © Deutsches Archaologisches Institut, 

replaced by open arches. 82 Palace interiors were embellished with paintings (only small 
fragments of which have survived) and with plaster panels decorated with animal images 
or ornamentation. Palaces invariably had gardens and parks, for which special irrigation 
canals were dug. 

The monumental architecture of the rule of Khusrau II (590-628) differs considerably 
from everything built earlier. The palace of Imarat-i Khusrau near Qasr-i Shirin, although 


Shepherd, 1983, p. 1067. 



ISBN 978-92-3-103211-0 Contents Sasanian religious architecture 

laid out according to the traditional design, was built on an artificial terrace 8 m high. 83 The 
palace in Bisutun was built on a stone platform reminiscent of Persepolis. Colonnades came 
into use and elements of Achaemenid architectural decoration were copied. The Ctesiphon 
palace (which is of burnt brick) differs from the traditional Sasanian design; built probably 
in the Early Sasanian period on the site of an old palace of the Parthian kings, it largely 
followed its predecessor's layout, which resembles that of the Parthian palace discovered in 
Ashur. The facade is composed of six rows of blind arches and half-columns. The throne 
room is roofed with a parabolic vault spanning 26.5 m. The interior includes five large 
rectangular rooms, a corridor and a number of small rooms. 84 

Sasanian religious architecture 

Zoroastrian religious structures of the Sasanian period are of two basic types: isolated 
square structures with circumambulatory corridors; and open-sided domed pavilions whose 
cupola stood on squinches resting on pillars at the corners that were linked by arches. 
Such shrines might serve as the centrepiece of a complex architectural ensemble. An Early 
Sasanian temple of the first type has been discovered at Bishapur. In ground plan it resem- 
bles the Parthian-period temple in Hatra. The square central building, each side of which 
was 14 m long, had four entrances. Another structure had a roof supported by Achaemenid- 
style imposts in the shape of the front part of bulls' protomes. The walls were built of 
rough-dressed blocks of sandstone without mortar. S5 Chahdr-taqs (simple domes set on 
four pillar-like walls) were common throughout Iran and images of them may be found 
on Sasanian vessels. 86 The ensemble of shrines near Firuzabad included both a chahdr-tdq 
and an Achaemenid-type fire-temple. 87 

At the site of Takht-i Sulaiman, a fire-temple has been discovered based on the same 
square plan with a domed roof, next to which were buildings for the priests, reception 
rooms and storehouses. A stone altar has been excavated similar to those shown on fifth- 
and sixth-century Sasanian coins. To the front of the temple complex, in the southern part 
of the site, there was a large lake. Together with the sacred fire, Anahita (the goddess of 
the heavenly waters) may well have been worshipped there. The ensemble dates from the 
fifth and sixth centuries. 88 There is an unusual temple, perhaps dedicated to Anahita, in 

83 Porada, 1965, p. 196. 

84 Reuter, 1930; Shepherd, 1983, p. 

85 Ghirshman, 1962a, p. 148. 

86 Naumann, 1977, p. 43. 

87 Vanden Berghe, 1961, p. 175. 

88 Naumann, 1977, p. 48. 




ISBN 978-92-3-103211-0 Contents Sasanian art and crafts 

Kangavar. It was probably raised on the site of an earlier temple to Artemis-Anahita that is 
mentioned by Isidore of Charax. The temple, which has not yet been fully excavated, stands 
on a stone platform 8 m high with a colonnade around the perimeter. Two gentle stone 
stairways lead to the platform. The structure has much in common with the architecture of 
the Achaemenid period but was probably built under Khusrau II. 89 

Sasanian art and crafts 

Early Sasanian art is of the proclamatory type. Designed to assert the divine nature of 
kingship and to reinforce the state religion, Zoroastrianism, it has much that is reminiscent 
of official Achaemenid art. The origins of the Sasanian pictorial canon appear to lie in the 
culture of pre-Sasanian Fars. Much, too, was drawn from Parthian culture: both Parthian 
investiture reliefs and inscriptions and the relief of the Sasanian shahanshahs stress the 
concept of legitimacy. Graeco-Roman cultural influence is discernible in several genres of 
Iranian art at various times. In the third century the 'western style' became widespread due 
to Shapur I's victories over the Roman emperors, the seizure of cities in Rome's eastern 
provinces and the migration of Greek, Syrian and Roman artisans and artists to Iran. At the 
turn of the fifth and sixth centuries there was a second wave of western influence; themes 
relating to the cult of Dionysus became common on precious metalware and glyptic, while 
Dionysiac scenes appear to have been included in the Zoroastrian festivity cycles. 90 

The proclamatory nature of Early Sasanian art is best seen on cliff carvings. The choice 
of topics here is severely restricted: investitures, triumph and duel scenes or portraits of the 
shahanshah and his courtiers. The rules for the representation of the king and courtiers that 
emerged in the first years of Ardashir I's rule are much the same as those for the last kings 
of the local Persian dynasty that preceded the Sasanians, as illustrated on their coins. 91 
The earliest known relief of Ardashir was carved in the ravine of Naqsh-i Rajab. Ardashir 
is shown standing before the god Ohrmazd (Ahura Mazda), who is crowning him king. 
Zoroastrianism personified its chief divinities at a relatively late period, and so Ohrmazd, 
the chief divinity, is portrayed in royal garb and crown. It was then, too, that the images 
of Anahita and Mithra were defined, as represented on coins of Hormizd I (271-272) and 
on reliefs of Narseh (293-303) and Ardashir II (379-383). A radiant crown was added to 
Mithra's kingly regalia while Anahita was shown dressed as the queen of queens. 

In another investiture relief both monarch and divinity are depicted mounted, with the 
vanquished Parthian emperor Ardavan V lying beneath the hooves of Ardashir I's horse and 

89 Lukonin, 1911b, p. 105. 

90 Lukonin, 1977a, p. 160. 

91 Lukonin, 1969, p. 23. 



ISBN 978-92-3-103211-0 


Sasanian art and crafts 

FIG. 9. Bahram IFs relief at Naqsh-i Rustam. (Courtesy of M. I. Mochiri.) 

the god of darkness, Ahriman, beneath those of Ohrmazd's steed (see Fig. 1). This graphic 
symbolism equates Ardashir's triumph with that of good over evil. 92 In the Firuzabad relief 
Ardashir Fs victory over Ardavan V is shown in a series of mounted duels, with Ardashir 
unhorsing the Parthian emperor and his heir Shapur, a Parthian grandee. 93 The poses of 
the steeds of both the defeated and their adversaries are reminiscent of images on Parthian 
reliefs such as those at Bisutun or Tang-i Sarvak in Elymais province. 94 The victories 
of Ardashir I and Shapur I over the Romans were reflected in reliefs showing scenes of 
triumph. 95 They were modelled on portrayals of the Roman emperors' triumphs: for all the 
differences in composition, many details are identical. 

Under Shapur I and Bahram II (275-293) (Fig. 9), reliefs were carved that depicted the 
royal court, with the figures of the king, the crown prince, the queen of queens and the 
courtiers being positioned in a strictly defined order. The Zoroastrian mobad (high priest) 
Kartir (Kirder) figures prominently in the reliefs of Bahram II. Four of Kartir's inscriptions 
are extant, containing his Creed and relating his activities. Alongside them, at Naqsh-i 
Rajab, there is a portrait of Kartir himself that is unique in Sasanian art (Fig. 10). 96 Also 
unusual is a relief showing Bahram II in single combat with lions at Sar Mashhad. 97 

92 Porada, 1965, p. 203. 

93 Ghirshman, 1962a, p. 125. 

94 Lukonin, 1977a, p. 146. 

95 Ghirshman, 1962a, p. 158. 

96 Lukonin, 1977a, p. 208. 

97 Ghirshman, 1962a, p. 173. 



ISBN 978-92-3-103211-0 


Sasanian art and crafts 

FIG. 10. Portrait of Kartir at Naqsh-i Rajab. (Photo: © Barbara Grunewald.) 

FIG. 11. Investiture relief of Narseh at Naqsh-i Rustam. (Courtesy of M. I. Mochiri.) 



ISBN 978-92-3-103211-0 


Sasanian art and crafts 

FIG. 12. Investiture relief of Ardashir II at Taq-i Bustan. (Photo: © Barbara Grunewald.) 

The shahanshahs who ruled after Bahram II have left only occasional reliefs, most of 
them investiture scenes. On the relief of Narseh at Naqsh-i Rustam he is crowned not by 
Ohrmazd but by Anahita, the patron of the Sasanian dynasty (Fig. 1 1). Apparently Narseh 
wished to stress the legitimacy of his claim to the throne he had seized and the end of 
Kartir's influence on the affairs of state. Ardashir II and Shapur III (383-388) each left 
one relief. The figures in fourth-century reliefs are more stylized and static; great attention 
was paid to decorative finish and to the representation of details of regal garb. Particularly 
interesting is Ardashir II's relief at Taq-i Bustan near Kermanshah (Fig. 12), 98 where the 
crown is given to the king by Ohrmazd who stands before him, while the god Mithra is 
shown standing on a lotus blossom behind Ardashir's back. Ohrmazd and Ardashir are 
trampling the figure of a fallen enemy (Fig. 13)." 

We know of no fifth- or sixth-century reliefs: the revival of the genre dates from the time 
of Khusrau II (590-628). The reliefs carved in the 'great grotto' at Taq-i Bustan during his 
reign differ from the earlier examples in both technique and content (Fig. 14). The central 
scene depicts an investiture - the king is shown in the centre, larger than the figures of the 
divinities. Below is probably the image of the shahanshah, mounted and in military array, 
executed almost as a sculpture in the round. 100 To the left and right, on the side walls of the 
niche, the king is shown hunting boar in a reedy marsh or gazelle and onager in a game park 

98 Ghirshman, 1962a, p. 190. 

99 Lukonin, 1969, p. 147. 
100 Porada, 1965, p. 207. 



ISBN 978-92-3-103211-0 


Sasanian art and crafts 

~ '&' 1 W 


FIG. 13. Detail of Ardashir IFs relief at Taq-i Bustan. (Photo: © Barbara Grunewald.) 

(Figs. 15 and 16). The scenes are executed in bas-relief with careful attention to detail. The 
image of the shahanshah is repeated several times. Taken together, the separate episodes 
make up a complete account that, in conjunction with the spatial aspect of the composition, 
is reminiscent of Assyrian reliefs. The arch-shaped facade of the niche is faced with stone 
blocks. In the centre are a crescent and ribbons towards which two winged Nikes are flying; 
the figures of the goddesses and the ornamentation at the edge of the arch reveal a strong 
Byzantine influence (Fig. 17). 101 

The work of Iranian metalsmiths is represented by dozens of examples in the major 
museum collections, testifying to the wide spread of these artefacts in antiquity. They 
include gold and silver cups, vessels, pitchers and rhytons executed in various techniques. 
Casting, engraving, embossing and crusta technique might be combined in one and the 
same object. The earliest Sasanian metalwork was decorated chiefly with portraits of the 
shahanshah, members of his family, princes or grandees. This tradition may date back to 
the Parthian period, when a famous medallion bearing the portrait of a Parthian emperor 
was set into the centre of a vessel. Early metal artefacts include a cylix from Sargveshi 
(Georgia) bearing the portrait of Bahrain II, Queen Shapurdukhtak and the Sakanshah 


Lukonin, 1977a, p. 187; Fukai and Horiuchi, 1972, PI. XXI et seq. 



ISBN 978-92-3-103211-0 


Sasanian art and crafts 

FIG. 14. The 'great grotto' at Taq-i Bustan. (Photo: © Barbara Grunewald.) 

(king of the Sakas) Bahram, and a cup bearing the portrait and inscription of the bitakhsh 
(prince) Papak. 102 

The subject that is most characteristic of Sasanian vessels, the shahanshah hunting 
on horseback, was only beginning to develop; it became widespread on fourth- and fifth- 
century artefacts. The king or crown prince is represented hunting lion, boar, ram or ante- 
lope. There are generally an even number of animals - two or four. The hunter attacks them 
with a bow or, more rarely, with a sword (Fig. 18). There are also images of kings hunting 
on foot; one example is a vessel bearing the portrait of Peroz (457-484) shooting caprid. 


Lukonin, 1979, p. 35. 



ISBN 978-92-3-103211-0 


Sasanian art and crafts 

FIG. 15. Detail of a royal hunting scene at Taq-i Bustan. (Photo: © Barbara Grunewald.) 



ISBN 978-92-3-103211-0 


Sasanian art and crafts 

" f f 






FIG. 16. Detail of a royal hunting scene at Taq-i Bustan. (Photo: © Barbara Grunewald.) 



ISBN 978-92-3-103211-0 


Sasanian art and crafts 

FIG. 17. Detail of ornamentation of the arch at Taq-i Bustan. (Photo: © Barbara Grunewald.) 

The border around the edge of the vessel is a barricade of nets behind which the heads of 
beaters and dogs may be seen. 103 Other subjects relating to the image of the king include 
Yazdgird II feasting with the queen, or Bahram Gur (Bahram V) and Azada. 

On sixth-century vessels, hunting scenes fade into the background - a vessel show- 
ing Khusrau I surrounded by his courtiers bears a royal hunting scene below the main 
composition. 104 Subjects from the Dionysiac cycle now became common - Sileni, mae- 
nads, theatrical masks and plant ornamentation - while scenes of the triumph of Diony- 
sus that are typical of Roman and Byzantine metalwork were copied. Such vessels were 
probably used during Zoroastrian festivals. 105 A characteristic group is composed of rhy- 
tons in the shape of animal heads, a form common in Achaemenid times and known to the 

103 Trever, 1937, p. 6. 

104 Porada, 1965, p. 217. 

105 Lukonin, 1977a, p. 160. 



ISBN 978-92-3-103211-0 


Sasanian art and crafts 







■ '^.- 

<r ' 






* [ ^^Jy 

FIG. 18. Silver dish showing the king on a lion hunt. Photo: © The Trustees of the British Museum. 

Parthians. Rhytons were used for ritual purposes while vessels bearing the king's portrait 
were traditional gifts from the shahanshah to his friends; many bore inscriptions showing 
the name of the owner and the weight. 106 

Little Sasanian jewellery is known. One of the most famous pieces is the so-called cup 
of Khusrau from the Abbey of Saint-Denis in France, which is decorated with coloured 
glass set in gold (Fig. 19). In its base is an inset disc of rock crystal carved in the image of 
an enthroned king. 107 In one hoard of coins in Iran an earring was found resembling those 
worn by Sasanian rulers as portrayed on coins. From Late Sasanian burials in Dailam come 
sword sheaths with gold and silver mountings decorated with filigree and granulation. 

Sasanian gems are almost as common as coins and thousands have been described from 
various collections. 108 The bulk of the gems used were semi-precious stones: chalcedony, 

106 Livshits and Lukonin, 1964, p. 155. 

107 Shepherd, 1983, p. 1102. 


Borisov and Lukonin, 1963; Bivar, 1969b; Gignoux, 1978. 



ISBN 978-92-3-103211-0 


Sasanian art and crafts 


m -m 

FIG. 19. Cup of Khusrau. Photo: © Bibliotheque Nationale de France. 

FIG. 20. Amethyst ring. Hermitage Museum, St Petersburg. (Photo: © Vladimir Terebenin.) 

amethyst, cornelian and lapis lazuli. These were either carved into ellipsoids and worn on 
laces or set into rings (Fig. 20). They were also used as seals. The most varied representa- 
tions can be found on them - portraits of kings, grandees and private individuals, horsemen, 
scenes of sacrifice or feasting, fabulous creatures, animals, birds and symbols or devices. 
The image was frequently accompanied by an inscription giving the owner's name and 
title or an auspicious Zoroastrian phrase. The subjects reveal the influence both of ancient 
Iranian traditions and of Graeco-Roman culture. One group is of seals of officials which 
bear only inscriptions. Another group is of items that belonged to Iranian Christians, on 
which the symbol of the cross is sometimes combined with a subject from the Zoroastrian 
cycle. Each period had its own range of subjects and stylistic features. The images were 
executed either by careful working of the detail or in line technique. The chronological 



ISBN 978-92-3-103211-0 Contents Sasanian art and crafts 

FIG. 21. Amethyst showing the portrait of 'Queen of Queens' Denak. Hermitage Museum, 
St Petersburg. (Photo: © Vladimir Terebenin.) 

classification of Sasanian carved gems is still far from complete. The best-known early 
gems are the British Museum's stone bearing the portrait of Bahram, king of Kerman; the 
Hermitage Museum's amethyst showing the portrait of 'Queen of Queens' Denak (Fig. 21); 
and the Bibliotheque Nationale's gem representing a horseman. Stones of the later period 
are chiefly cut in line technique. 

Sasanian ceramic vessels vary greatly in form and size. Each region had its own char- 
acteristic pottery linked with the previous period. The most common forms were vari- 
ous kinds of pitchers made on a quickly turning wheel. We know of vessels on which 
the potter's name was inscribed on the raw clay before firing. 109 The ornamentation was 
usually restricted to rows of straight or wavy horizontal lines, stamps or applique work. 
Throughout the Sasanian period, pottery was produced (rhytons and small, pear-shaped 
pitchers) that imitated metal artefacts. 110 Somewhat different are the Late Parthian and 
Early Sasanian vessels finished in green glaze that were common in Mesopotamia and 
Khuzistan. 111 Glassware is known from excavations in Kish, Ctesiphon and Susa and from 
Early Sasanian burials in northern Mesopotamia. The chief glass-making regions were 
Mesopotamia and north-western Iran. Sasanian glassware developed in isolation, although 
some forms are reminiscent of Syrian and Roman artefacts. Glass vessels were engraved; 
some extant examples are decorated with patterns cut from gold leaf and encased between 
two layers of glass. The most common forms were thick-walled spherical cups, with cut 
ornamentation made of slightly concave discs or oval facets (Fig. 22). Glass rhytons and 
amphorae were also produced. 112 

109 Huff, 1978, p. 145. 

110 Ettinghausen, 1938, pp. 664 et seq. 

111 Riccardi, 1967, p. 93. 

112 Fukai, 1977. 



ISBN 978-92-3-103211-0 Contents Sasanian art and crafts 

FIG. 22. Thick-walled spherical cups. (Photo: © Vladimir Terebenin.) 

Weaving was a highly developed craft. Samples of Iranian silk cloth were kept in the 
church treasuries of Western Europe, and fragments have been found in Egyptian and Cen- 
tral Asian burials. Cloth was woven in the compound twill technique and was brightly 
coloured. The most common pattern was one of rhombic or round medallions with pic- 
tures of the senmurws, winged horses, animals or birds, placed either in vertical rows or 
at random. An idea of the ornamentation and colouring of cloth and the cut of the cloth- 
ing may be gained from Central Asian paintings or the Taq-i Bustan reliefs, which show 
the costumes of the king and courtiers. A royal hunting costume designed for wearing on 
horseback, and consisting of trousers and a short tunic, is depicted on Sasanian vessels. 
Iranian clothing is described by Ammianus Marcellinus (XXIII, 6.84): 'Most of them are 
so covered with clothes gleaming with many shimmering colours, that although they leave 
their robes open in front and on the sides, and let them flutter in the wind, yet from their 
head to their shoes no part of the body is seen uncovered.' The products of Iranian weavers 
were in demand in many countries, where they were considered luxury items. The patterns 
on Sasanian textiles were long copied virtually unchanged in Byzantium and were imitated 
in Central Asia and China. 113 The artistic influence of Sasanian Iran may be detected in 
the cultures of many countries, from Western Europe to Eastern Asia. It made a major 
contribution to the subsequent development of the visual culture of the Muslim East. 

113 Shepherd, 1983, p. 1107. 



ISBN 978-92-3-103211-0 


Sasanian art and crafts 

Table 1. The Sasanian Empire 1 (courtesy of M. I. Mochiri) 

Pajjak (2C*-222) 

(DAdashirl (224-24 1) 

(2>SWir I {241-271) 

. !- , 

(J) Hoimizd I (171-372) (7) Narwh (293- 3-03> (4) Bahram [ (272-275) 

I I 

(B> HomuuJ 11 (301-309) (5) Bah«m 11 (275-293) 

i \ I 

< I ty ArtLuhif II {379-38 3) (9) Shipur 1 1 (309-379) ft) B*h*iP III (293) 


(16) Hmrmizd III (4*7-459) 

{11 J Shipm lEJJU-JJU) (!2}H>hMjn[V(3ftlt-399) 


(13) Yazdgird J (399-42Q) 


(14) V (420-438) 

(15) Yjttdgird II (436-457) 





(l8)Vj| M h(4S+~l3$) 


( 19) Kjivad I (4BS-t97> 

and (499-5:51) 



(22) Hc.frt.i2i IV (S79-590) 

(20) Jairusp (497-*9?) 


{23) Bahmm VI (590-591) 

(24) Viscahm (591-597) 

(25) Hormizd V {}5»>* 



(29) Khuirau III 

(30) &w, n {629-630) 

(32) Hannizd Vt {631-632) 1 

(27) Kav*d II (623) 

[ZS) Ardashir III {&ZS-629J 

{63l) f «*L«« 


) YaraJginJ 

1. For some alternative dates, see Frye, 1984, p. 361. (Note, however, that some authors in this 
volume have their own datings.) 

2. Mochiri, 1977, pp. 209-15. 

3. Mochiri, 1972, pp. 17-18; 1977, p. 203; 1983, pp. 221-3. 

4. Mochiri, 1972, pp. 11-16; 1977, pp. 203-5. 

5. Mochiri, 1972, pp. 13-16; 1977, pp. 205-8. 

6. Mochiri, 1977, pp. 174-202. 



ISBN 978-92-3-103211-0 Contents SASAN IAN IRAN 


A. Tafazzoli and A. L. Khromov 


Written works 82 

Secular literature 82 

Royal Res Gestae 86 

Religious literature 88 

Philosophy and theology 90 

Visionary and apocalyptic texts 90 

Andarz (wisdom literature) 91 

Languages and scripts 91 


Borrowings and influences 94 

Science and philosophy 96 

Court chronicles and epic histories 99 

Religious life 102 

Religion and the law 105 

See Map 2. 



ISBN 978-92-3-103211-0 Contents Secular literature 

Part One 
Written works 

(A Tafazzoli) 

Literary works extant from the Sasanian period may be divided into two groups, reli- 
gious and secular. As the secular literature was written within the framework of Zoroastrian 
religious beliefs, however, it also manifests religious overtones. Translations of, and com- 
mentaries upon, the Avesta (the sacred book of the Zoroastrians) in Middle Persian (also 
known as Pahlavi), as well as books written on the basis of oral traditions of Avestan mate- 
rial, constitute the most important of the religious works. Literature of the Sasanian era 
bears the characteristics of oral literatures. 

Secular literature 

The interest in oral literature in pre-Islamic Iran 1 meant that, apart from state or commercial 
records and documents and, on rare occasions, religious works, nothing was written down 
until the Sasanian period. Secular literature was preserved orally by gosdn (poet-minstrels) 
or khunyagar 2 (story-tellers). When Middle Persian had become obsolete and the religion, 
rituals and customs of the Iranians had undergone changes, the originals of many literary 
works of this type were lost. Thus our information, especially on the secular literature of 
this period, is based on secondary sources. 


The core of Iranian epic stories belonged to the Avestan people of eastern Iran. We find ref- 
erences in the Avesta to its heroes, especially the Kayanian princes who were the ancestors 
of Gushtasp (an Iranian king and the patron of Zoroaster). These stories, which recounted 
the deeds of military commanders and heroes of old, were gradually transformed in the 

1 On the oral tradition, see Bailey, 1943, pp. 149 et seq; Boyce, 1968b, pp. 32 et seq. 

2 Boyce, 1957, pp. 10-45; Tafazzoli, 1968, pp. 410-11. 



ISBN 978-92-3-103211-0 Contents Secular literature 

minds of the people into marvellous feats, accumulated a wealth of detail, and were handed 
down orally - frequently in versified form - to later generations. There are three cycles of 
traditions concerning heroic tales in the Iranian national epic: the Kayanian, the Saka and 
the Parthian cycles. 

Epic stories, frequently in verse, remained an oral form until the Sasanian period and 
some were used in the compilation of the Khwaddy-ndmag [Book of Lords] (see page 84) 
in Pahlavi. The only extant work of this type is the versified Ayadgar-T Zareran [Memoirs 
of Zarer's Family]. This work was originally in the Parthian language and found its final 
redaction in a summarized, written form, probably towards the end of the Sasanian era. It 
concerns the wars between the Iranians and the Turanians after the conversion to Zoroastri- 
anism of Gushtasp, the Iranian king. Zarer, the king's brother, was slain during these wars. 
The Parthian words appearing in the text betray its Parthian origin. Its poetic language is 
clear. A more detailed version of the story of Zarer appears in the tenth-century Shah-name 
[Book of Kings] of Firdausi, who quotes it from Daqiqi. 3 

Titles of other epic stories, which probably existed independently of the Khwaddy- 
ndmag and were later translated into Arabic, are mentioned in Islamic sources. Examples 
are: The Story ofRustam and Isfandiydr; 4 the Sagesardn [Leaders of the Sakas]; 5 and the 
book of Paykdr on the battles of Isfandiyar. 6 Short pieces of lyrical, panegyric or other 
types of poetry are also found in Persian and Arabic books from the early Islamic era. 7 


Towards the end of the Sasanian period, especially during the reign of Khusrau I (53 1-579) 
and later, increasing attention was paid to the task of collecting legends. The original ver- 
sions in Pahlavi have been lost, but on the basis of their Arabic and Persian translations, 
as well as references made to them in Islamic sources, the books of stories of the Sasani- 
ans appear to fall into two groups: Iranian tales; and tales adapted or translated from other 
languages into Middle Persian. 

The most important collection of Iranian tales was the Hazdr afsdn [The Thousand 
Tales], mentioned by both Ibn al-Nadim 8 and al-Mas c udi. 9 This work was translated into 
Arabic, enjoyed widespread fame among Muslims, and was used as the basis for the 

3 Benveniste, 1932; Utas, 1975. 

4 Ibn al-Nadim, 1973, p. 364. 

5 Al-Mas c udi, 1965, Vol. 1, p. 267; Christensen, 1932, pp. 142 et seq. 

6 Al-Mas c udi, 1965, Vol. 1, p. 229; Hamza, 1967, p. 94; Christensen, 1932, pp. 143 et seq. 

7 Tafazzoli, 1974, pp. 338, 347-9. 

8 Ibn al-Nadim, 1973, p. 363. 

9 Al-Mas c udi, 1965, Vol. 2, p. 406. 



ISBN 978-92-3-103211-0 Contents Secular literature 

compilation of The Thousand and One Nights. The present Arabic version, dating back 
to the fourteenth century, contains, in addition to Persian stories, others which were preva- 
lent in Baghdad and Egypt at various periods. In addition to the Hazar afsan, Ibn al-Nadim 
mentions a book entitled the Hazar dastan 10 [The Thousand Stories]. 

The most famous collection of tales translated into Pahlavi was the Kalilag u Dimnag 
[Kalila and Dimna], the original source for which was the Indian Panchatantra, probably 
its Prakrit version. 11 This work, which is now lost, was translated into Pahlavi in the middle 
of the sixth century by the physician Burzoe for Khusrau I (see also page 94). Ibn al- 
Muqaffa c and c Abdallah Ahwazi then translated it into Arabic. A Syriac translation of the 
Pahlavi text, dating from 570, is also extant. The Kalilag u Dimnag was also translated 
into Latin and Greek, and from Arabic into Persian on several occasions. The earliest of 
these Persian translations, dating back to the first half of the tenth century, has been lost; 
the most famous is the twelfth-century version by Abu al-Ma c ali Nasrallah Munshi. 

Another work of Indian origin which was compiled in Iran is the Sindbad-namag [Book 
of Sindbad], which probably dates back to the time of Khusrau I. Like the Kalilag u Dim- 
nag, this work was also translated into Arabic, probably by the ninth-century translator 
Musa b. c Isa al-Kisrawi. According to Ibn al-Nadim, 12 two prose versions existed in Ara- 
bic. The Sindbad-namag was then translated into Persian in the tenth century and was later 
versified. Neither version has survived; the only available translation in Persian dates from 
the twelfth to the thirteenth century. It owes its fame in Europe to the Greek and Latin 

The Bilauhar u Budasaf is another story of Indian origin which was translated into 
Pahlavi during the Sasanian period, and thence into Arabic in the Islamic era. This work, 
which is basically an account of the Buddha's life, is not a translation of a specific Indian 
text but, rather, a collection of the legends surrounding the life of the Buddha. Its earliest 
translation into Arabic dates back to the eighth century; a tenth-century translation is also 
extant. Parts of the Bilauhar u Budasaf 'in classical Persian verse, written in Manichaean 
script, have been found amongst the Turfan fragments belonging to the followers of Mani. 13 
The work was translated in the eleventh century into Georgian, Greek and Latin and, 
later, into other European languages. Many traces of the Bilauhar u Budasaf are found 
in Persian and Arabic books of manners and ethics. 14 The TutT-namag, a translation of and 

10 Ibn al-Nadim, 1973, p. 364. 

11 Mojtabayi, 1984. 

12 Ibn al-Nadim, 1973, pp. 186, 364. On the Sindbad-namag, see Minovi, 1968, pp. 169 et seq. 

13 Henning, 1962, pp. 91 et seq. 

14 Gimaret, 1971; Sundermann, 1982, pp. 101 et seq. 



ISBN 978-92-3-103211-0 Contents Secular literature 

adaptation from the Indian Sukasaptati [Seventy Tales of the Parrot], several Persian ver- 
sions of which exist, was probably translated into Pahlavi in the Sasanian period. 

Among the romantic tales in Pahlavi, the most important is the Vis u Rdmin, which is 
of Parthian origin and was translated into Pahlavi towards the end of the Sasanian period 
and into Persian in the Islamic era. (Both the Parthian original and the Pahlavi translation 
are lost.) Fakhr al-Din As c ad Gurgani turned it into Persian poetry early in the eleventh 
century. The work contains important information on the life, manners and customs of the 
Parthians. 15 

The VdmTq u c Adhra is a romantic story of Greek origin which is said to have been 
translated into Pahlavi at the time of Khusrau I. 16 It was later translated from Pahlavi into 
Arabic (both these versions are lost). A Persian translation was made (probably from the 
Arabic) which was then turned by c Unsuri (eleventh century) into poetry - some verses are 
extant. Part of it is also quoted in the Ddrdb-ndme} 1 a Persian tale by the twelfth-century 
writer Tarsusi. 

The Alexander Romance by pseudo-Callisthenes may have been translated into Pahlavi 
towards the end of the Sasanian period, probably from the Greek. It appears that a Syr- 
iac translation was later made by Christians living in Iran and using oral as well as writ- 
ten sources. Arabic translations were made from the Pahlavi and Syriac versions of the 
Iskandar-ndme. Versions of this romance also exist in Persian, the earliest belonging to 
some time between the twelfth and the fourteenth century 18 - in addition to the Arabic 
text, the authors used oral traditions. 

The Kdrndmag-T ArdasTr-T Pdbagdn [Book of the Deeds of Ardashir Papakan], in 
Pahlavi, belongs to the epic genre (see also page 96). Its present version dates back to 
the Late Sasanian period and contains certain reworkings from Islamic times. The biog- 
raphy of Ardashir I (226-241) is narrated in this work through a mixture of legend and 
historical fact. 


There were several handbooks in Pahlavi dealing with institutions, court manners and 
ceremonies, the duties of the various social classes, the rules of battle, the arts of war- 
fare (horsemanship and shooting), and games and entertainments (such as polo, chess and 

15 Minorsky, 1964. 

16 On the VdmTq u c Adhra, see Shafi c 's edition, 1967, Introduction. 

17 Safa, 1965, Vol. l,p. 209. 

18 Afshar, 1964, Introduction, pp. 9-37. 



ISBN 978-92-3-103211-0 Contents Royal Res Gestae 

backgammon). The originals of these Ay en Ewen-ndmag are lost, but some were translated 
into Arabic, parts of which, or references to them, are available in Islamic sources. 19 


The Mddlgdn-T hazdr dddestdn [Book of a Thousand Judicial Decisions] (see also pages 
101-2), compiled by Farrukhmard, the son of Bahram, is the most important collection of 
legal texts of the Sasanian period. This work, which contains a number of legal cases con- 
cerning marriage, inheritance, ownership, endowments, and so on, was probably compiled 
at the time of Khusrau II, but its final redaction belongs to the ninth century. 20 Zoroastrian 
religious laws are also collected in other works. 21 In the field of education, a short extant 
treatise entitled the Khweskdrlh-i Redagdn [The Duties of Children] contains instructions 
concerning children's obligations towards their parents and teachers. 22 

The sole surviving Pahlavi work on geography is the Sahnhd-l Erdn, a treatise contain- 
ing a list of the major cities and fire-temples of Iran. Although it was written in the ninth 
century, its contents are mostly mythical and relate to ancient times. 23 

Royal Res Gestae 

Important events of the reign of each of the Sasanian kings were written down and pre- 
served in the imperial archives, a practice that probably dates from the very beginning of 
Sasanian rule. Shapur I (241-271) left a description of his deeds in a trilingual inscription 
on the wall of the Ka c be of Zoroaster. Narseh (293-303) left an account of his accession to 
the throne in a bilingual inscription at Paikuli (situated in the Zagros mountains in present- 
day Iraq). Copies of such works were probably preserved in the imperial archives. Through 
his learned Syrian friend Sergius, the sixth-century Greek historian Agathias gained access 
to the official written documents of the time of Khusrau I which were kept in the impe- 
rial archives. References in the Shdh-ndme of Firdausi suggest that when Hormizd IV 
(579-590) was imprisoned by his general Bahram Chobin and the nobility in c. 590, he 
expressed the desire for someone to come with a book and read the stories of the past kings 

19 Christensen, 1944, pp. 61-2; Tafazzoli, 1976, p. 266; 1985, p. 692. 

20 Perikhanyan, 1973; Macuch, 1981. 

21 Tavadia, 1930; 1956; Kotwal, 1969; Safa-Isfehani, 1980; Anklesaria, 1960; Jamaspasa and Humbach, 

22 Junker, 1912a; Freiman, 1918. 

23 Markwart, 1931. 



ISBN 978-92-3-103211-0 Contents Royal Res Gestae 

to him. 24 Islamic sources mention the existence in the Sasanian period of official registers 
in which the events of each king's reign were registered along with his portrait. 25 


The idea of compiling a written national history for the Iranians appeared towards the end 
of the Sasanian period, especially at the time of Khusrau I, during whose reign books were 
either written in Pahlavi or translated from other languages, such as Syriac, the Indian 
languages and Greek. The Khwdddy-ndmag was probably completed during the reign of 
Yazdgird III, the last of the Sasanian kings, who assumed power towards the end of 631 or 
the early part of 632. 26 The work detailed events from the creation of the world up to the 
end of the reign of Khusrau II (590-628) and conformed to the viewpoint of the Zoroastrian 

The Pahlavi original of this history was lost, but Arabic translations were made in 
Islamic times. The translators did not limit themselves to a literal translation of the orig- 
inal. The oldest translation is that of Ibn al-Muqaffa c (c. 720- 756). 27 Although none of 
these Arabic translations has survived intact, they served as the basis for the history of the 
pre-Islamic period in the works of the Muslim historiographers. The Khwdddy-ndmag was 
turned into Persian prose and verse in the tenth century. Among the prose versions, the 
most famous is the Shdh-ndme of Abu Mansur, completed in 957. Both the Arabic trans- 
lations and the Pahlavi sources were used in the compilation of this Shdh-ndme, which, 
in turn, was used along with other oral and written sources by Firdausi in the creation of 
his epic. 

Religious traditions as they existed in the Avesta of the Sasanian period, as well as 
its translations and the zand (commentaries upon it), naturally formed the basis of the 
Khwdddy-ndmag. According to Zoroastrian beliefs, the duration of the material world 
(gumekhtagTh, or 'mixture') is 6,000 years, divided into 6 periods each lasting 1,000 years. 
All the events and reigns of the kings were placed within this framework. As may be seen 
in the case of the Kdrndmagl Ardaslr-l Pdbagdn, the history of the fifth dynasty (i.e. the 
Sasanians), though partly based on court archives, was nevertheless influenced by legend. 

24 Noldeke, 1920, §12. 

25 Christensen, 1932, pp. 66-7. 

26 Noldeke, 1920, §13. 

27 On the Arabic translations of the Khwdddy-ndmag, see Christensen, 1925, pp. 23 et seq.; 1934, pp. 81 
et seq.; Osmanov, 1975, pp. 287 et seq. 



ISBN 978-92-3-103211-0 Contents Religious literature 


Several works discussed government policies and ways and means of governing the king- 
dom. Among them is the Name-TTansar [Letter of Tansar], written by Tansar (or, in the cor- 
rect form, Tosar), the Zoroastrian mobad (high priest) at the time of Ardashir I, in response 
to Gushnasp, king of Tabaristan. Although the original probably belonged to the time of 
Ardashir, changes were made to it in later periods, particularly during the reign of Khus- 
rau I. The Arabic translation by Ibn al-Muqaffa c is also lost and what survives is an early 
thirteenth-century re-translation into Persian from Ibn al-Muqaffa c 's version. This version 
is by Ibn Isfandiyar, who includes it in his History of Tabaristan. 29, 

The c AhdArdasir is another tract that bears a striking resemblance to the Name-TTansar. 
The original in Pahlavi is lost and it survives in an Arabic translation. 29 The Karnamag-T 
Anosirvan [Book of the Deeds of Anushirvan, i.e. Khusrau I] (see also pages 96-7), an 
account of the work undertaken during Khusrau's reign together with methods of adminis- 
tering the affairs of state, is also extant only in an Arabic translation. 30 

Religious literature 


The Avesta was preserved orally until towards the end of the Sasanian period, probably dur- 
ing the reign of Khusrau I, when it was compiled in 21 sections. To transcribe it, a script 
was invented from the consisting of about 46 characters. An attempt was made to record 
the Avestan words exactly as they were pronounced at the time. 31 Although the Avesta was 
now written down, because of the importance attached to the oral tradition, manuscripts 
of the work were rare and the recorded text was consulted infrequently. The oldest extant 
manuscript of the Avesta dates back to 1258 or 1278. In the Sasanian period, Avestan was 
considered a dead language. So that the contents of the Avesta could be understood, it 
was translated into Pahlavi and the zand (commentaries) were written upon it. Oral tradi- 
tions, as well as the sciences of the time - known to Iranians through the Greeks, Romans, 
Syrians and Indians - were also used in writing these commentaries. Today, translations 
of early works such as the Gathas, Yasnas, VendTdad, Nerangestan, Niyayish, AfrTnagan, 
Herbadestan and some parts of the Yasts are extant, but the translation of, and commentary 
upon, the complete Avesta certainly existed during the Sasanian period. 

28 Minovi, 1975; Boyce, 1968c. 

29 c AhdArdasir, 1967; Grignaschi, 1967. 

30 Grignaschi, 1967, pp. 16-45. 

31 Bailey, 1943, pp. 149-94. 



ISBN 978-92-3-103211-0 Contents Religious literature 

Based on the Avesta and its translations, many books and treatises were written on var- 
ious subjects some of which, like the Denkard, constitute a religious encyclopaedia. The 
contents of this book, which is essentially a compilation, belong to different periods. The 
final redaction took place in the ninth century and we know the names of two of its com- 
pilers: Adurfarnbag, son of Farrukhzad, and Adurbad, son of Emed. Adurfarnbag was the 
mobad of the Zoroastrians of Fars in this period. The Denkard was originally in nine books 
of which the first, second and part of the third were lost. The third book is nevertheless the 
largest; it deals with Zoroastrian religious principles, at times in a philosophical language, 
and at others in one mixed with mythical elements. The seventh book is a legendary biog- 
raphy of Zoroaster, based on a section of the Avesta called the Spand nask. The contents of 
the Sasanian Avesta in a condensed form constitute the eighth book. Since both the origi- 
nal and the Pahlavi translation of most of these parts are lost, these summaries are of great 
importance. The ninth book contains a commentary on three parts of the Sasanian Avesta 
which is particularly significant for its mythical contents. 32 

The Bundahisn [The Original Creation], an important work in Pahlavi, is also known 
as the Zand-agahTh. Its subject-matter ranges from cosmology, astronomy and eschatology 
to lists of rivers, mountains and plants. Although the final redaction belongs to the ninth 
century, it was probably compiled in the Late Sasanian period. Two forms are available: 
the more detailed is called the Great Bundahisn; and the shorter, the Indian Bundahisn? 3 

The Vizidagiha-l Zadsparam [The Selections of Zadsparam] is another important work 
that is similar to the Bundahisn in content and whose author has apparently used the same 
sources. Its compiler, Zadsparam, was a priest who lived in the ninth century. The book 
consists of four parts and deals with creation, cosmology and resurrection. It contains a 
section on the life of Zoroaster which is similar to the seventh book of the Denkard and 
uses the same sources. 34 

A book and three letters written by Manuchihr, leader of the Zoroastrians of Fars and 
Kerman, are extant. The book, entitled the Dadestan-l denig [Religious Judgments], con- 
tains 92 questions posed in written form by a Zoroastrian; it deals with a variety of topics 
such as the principles of Zoroastrian belief, cosmology, mythology and religious laws. 35 

The Pahlavi Rivayat is a collection of miscellaneous material probably compiled by a 
tenth-century author. The work is particularly significant as it contains certain myths which 
are not mentioned elsewhere or are mentioned in passing, as well as providing information 
on religious rituals and ceremonies and popular beliefs. The practice of collecting rivayat 

32 See Menasce, de, 1958; 1973, Introduction. 

33 Anklesaria, 1908; West, 1880, Introduction. 

34 Anklesaria, 1964, Introduction; Gignoux and Tafazzoli, 1993, Introduction. 

35 West, 1882, Introduction; Dhabhar, 1912, Introduction. 



ISBN 978-92-3-10321 1-0 Contents Visionary and apocalyptic texts 

(traditions) continued among the Zoroastrians for many centuries, a collection also being 
extant in Persian. 36 

Philosophy and theology 

Iranians were familiar with Greek philosophy from the Achaemenid period. This acquain- 
tance was deepened in Sasanian times, leading to the influence of Greek philosophy on 
Zoroastrian religious works. Although no philosophical works in Pahlavi are available 
from the Sasanian period, those written in the ninth and tenth centuries on philosophy 
and theology show that they are based on an older tradition. The third and fourth books 
of the Denkard, mentioned above, are among the principal philosophical and theologi- 
cal works. Another important example is the Skand gumdnlk wizdr [The Doubt-crushing 
Explanation], written by Mardan Farrokh, son of Ohrmazd-dad, towards the end of the 
ninth or early tenth century. 37 The Pus-l ddnisn-kdmag 3 ^ is a short treatise containing argu- 
ments similar to those found in the Skand gumdnlk wizdr. Another dialectical treatise is the 
Gujastag Abdlis [The Cursed Abalish]. It concerns a debate conducted in the presence 
of the c Abbasid caliph al-Ma'mun (813-833) between Abalish, a Zoroastrian converted 
to Islam, and Adurfarnbag, one of the compilers of the Denkard. The treatise appears in 
question-and answer form. 39 

Visionary and apocalyptic texts 

Visionary and apocalyptic literature goes back to the Avesta, but the oldest extant Pahlavi 
text in this genre is the Kartir (Kirder) inscription at Sar Mashhad, in part of which Kartir 
describes his ascension to the other world. Unfortunately, this text is fragmentary and dam- 
aged. Several Pahlavi texts are extant in this field, the most important being the Ardavlraz- 
namag [Book of Ardawiraz]. It concerns the ascension of Viraz (Viraza in Avestan) who, 
according to some scholars, is identified in the work with Vehshapur, a famous priest of the 
time of Khusrau I. Having taken an intoxicating substance, he travels to the other world, 
sees paradise, hell, purgatory, the rewards accorded to the pious and the punishments meted 
out to evil-doers, all of which he describes upon his return. Though the essential core of the 
book's contents is very old, the extant version dates from the tenth century or even a little 

36 Dhabhar, 1913, Introduction; Williams, 1990, Introduction. 

37 Menasce, de, 1945. 

38 Junker, 1959. 

39 Tafazzoli, 1985, p. 58. 



ISBN 978-92-3-103211-0 Contents Languages and scripts 

later. 40 Similar works in which the events of the world are foretold are the Zand-T Vahman 
yast 4i and the Jdmdsp-ndmag, or Ayddgdr-T Jdmdspig. 42 

Andarz (wisdom literature) 

An important section of Pahlavi literature consists of the andarz (wisdom) collections, 
which are of two types: religious advice and pragmatic wisdom. The subject-matter of most 
of the andarz pieces of the second type may be found in the literature of other nations too, 
and translations of this group later found their way into the books of adab ( literature) and 
ethics in Persian and Arabic. For those items which belong to the realm of oral literature, 
it is impossible to establish an author or a date. In Pahlavi, most such pieces are attributed 
to great and learned men in general - the Andarz-T ddnagdn u Mazdaesndn is one example. 
However, some items are attributed to specific kings, dignitaries or religious personalities. 
In the third book of the Denkard, the mythical Jamshid, the Peshdadian king, is credited 
with the authorship of a number of these. A collection is also recorded in the name of 
Khusrau I. Andarz collections may also contain pieces of poetry; some items are in metrical 
prose while others are merely endowed with a poetic quality. The most extensive of the 
andarz collections is the sixth book of the Denkard. 43 

Languages and scripts 

The most widespread languages during the Sasanian era were Middle Persian (or Pahlavi), 
Parthian, Sogdian, Khwarizmian, Khotanese Saka and Bactrian; various texts in these 
languages are extant. 


This is the development of Old Persian, or of one of its dialects, which used to be the 
language of the region of Fars. The evolution of Old Persian into Middle Persian probably 
began during the fourth century B.C., but the oldest extant documents in the latter language 
belong to the third century a.d. Middle Persian lacks gender and the dual form. With rare 
exceptions, the declensional forms of nouns and pronouns have disappeared. Among the 
items extant in Middle Persian are coins of the local kings of Persia (from about the middle 
of the third century B.C.), inscriptions by the Sasanian kings and dignitaries (mostly from 

40 Gignoux, 1984; Vahman, 1986. 

41 Anklesaria, 1957. 

42 Modi, 1903; Messina, 1939. 

43 Shaked, 1987, pp. 11-16. 



ISBN 978-92-3-103211-0 Contents Languages and scripts 

the third and fourth centuries a.d.), Sasanian coins and seals, Zoroastrian Pahlavi texts, an 
excerpt from the Psalms and Manichaean writings. Except for the last-mentioned, they are 
all written in different variants of Pahlavi script. 

Since the Pahlavi script was difficult, lists of words were prepared to be memorized 
by the scribes. The Frahang-T Pahlawig is one such work in which words with different 
spellings, as well as heterograms (incorrect spellings), were arranged systematically to 
make them easier to learn. 44 In this treatise, words are grouped under subject headings. An 
alphabetical version of the Frahang-T Pahlawig is extant in an eighteenth-century copy. 45 
A sheet from a Pahlavi glossary of the ninth or tenth century, containing seven Pahlavi 
verbs in the form of heterograms, was found amongst the Manichaean fragments discov- 
ered in Turfan. 46 An Avestan-Pahlavi glossary, entitled the Frahang-T (Dim Ewak, is extant; 
it is of special importance as it contains Avestan words used in the Sasanian period but 
subsequently lost. 47 

The Pahlavi script is divided into two main forms, the lapidary and the cursive, both of 
which derive from the Aramaic. In these scripts, written from right to left like the Aramaic 
prototype, the vowels are not usually represented. The lapidary script is generally used 
in inscriptions, coins and seals - though a variant of the cursive called kastaj or kashtaj 
(from the Pahlavi gastag) by writers of the Islamic era 48 is also sometimes used for these 
purposes. Unlike the cursive, the lapidary script includes 19 characters which cannot be 
joined. Middle Persian royal and private inscriptions of the Sasanian epoch are written in 
the lapidary script, the oldest being the short inscription of Ardashir I at Naqsh-i Rustam. 
Other important inscriptions are that of Shapur I on the wall of the Kacbe of Zoroaster (in 
Middle Persian, Parthian and Greek) and that of Narseh (dating from 293-294) at Paikuli. 49 
Of the private inscriptions the most important are four by Kartir, all of which are located in 
Fars. 50 The latest inscription of this type is that of Mihrnarseh at Firuzabad in Fars, which 
belongs to the first half of the fifth century. 51 

The Pahlavi cursive or book script contains 13 characters and was mostly used in writing 
Zoroastrian works in Pahlavi. In this script, one character often represents several sounds; 
moreover, a ligature may be read in different ways, so that the script could be ambiguous. 

44 Junker, 1912/?; Ny berg, 1988. 

45 Nyberg, 1970, pp. 343 et seq. 

46 Barr, 1936. 

47 Kligenschmitt, 1968. 

48 Ibn al-Nadim, 1973, p. 15; Hamza, 1967, p. 65. 

49 Humbach and Skjaerv0, 1978-83, p. 83. 

50 For the inscriptions of Kartir, see Gignoux, 1973; 1991. 

51 For the bibliography of the Pahlavi inscriptions, see Gignoux, 1972, Introduction, pp. 9-14. 



ISBN 978-92-3-103211-0 Contents Languages and scripts 

The oldest extant text in Pahlavi cursive was copied in 1323. 52 Cursive script was also 
used for writing on parchment, papyri, ostraca, stones and gems and it is found on tomb 
and funerary inscriptions like the Istanbul inscription. 53 Others of this type, and belong- 
ing to the Late Sasanian period, have been discovered in Fars. 54 The most recent funerary 
inscription, written vertically, is a ninth-century inscription found in China. 55 Commem- 
orative inscriptions include those at Darband (in the Caucasus) from the sixth century, 56 
the inscriptions on Shapur I's horse at Naqsh-i Rustam and Bishapur, as well as the two 
discovered at Maqsudabad and Tang-i Khushk and which include a statement explaining 
the ownership of the estate and its improvements. Finally, in cursive script, there are the 
tomb inscriptions of the rulers of Mazandaran (ninth-eleventh centuries). 57 A fragment of 
the translation of the Pahlavi Psalter, found in Central Asia, is in a variant of the cursive 
script. Its compilation dates back to the sixth century or earlier, but it was written down 
between the seventh and the eighth centuries or even later. 58 


Manichaean works are written in a script which is a variant of the Syriac script and is 
peculiar to the Manichaeans (see also pages 99-101). Twenty-two letters were adopted 
from the Semitic alphabet and a newly evolved letter - j - was added to it. This script 
underwent little evolution from the third century until it was abandoned, probably in the 
thirteenth century. 59 Manichaean works in Iranian languages - Middle Persian, Parthian, 
Sogdian and, later, Persian - were generally written in this script. In contrast to other 
systems, each character in the Manichaean script represents a single sound, and there are 
almost no historical or pseudo-historical spellings, nor are there heterograms. 


The Parthian language was spoken in the south-western areas of Central Asia and in 
Khurasan during the Arsacid period (third century B.C. - third century a.d.). It was a living 
language until some time in the Sasanian period. The main differences between Middle 

52 Henning, 1958, pp. 46-7; Boyce, 19686, p. 65. 

53 Menasce, de, 1967; Harmatta, 1969, pp. 255-76. 

54 Gignoux, 1972; Gropp, 1969; 1970; Harmatta, 1973, pp. 68-79; Tafazzoli, 1991. 

55 Harmatta, 1971, pp. 113-14; Humbach and Wang, 1988. 

56 Gropp, 1975; Kasumova, 1987; Gignoux, 1992. 

57 Henning, 1958, p. 50. 

58 Ibid., p. 47. 

59 Ibid., pp. 73 et seq. 



ISBN 978-92-3-103211-0 Contents Borrowings and influences 

Persian and Parthian lie in their phonology and vocabulary, their verbal systems demon- 
strating remarkable similarity. 

The principal remnants of the Parthian language include the ostraca (from between 100 
and 29 B.C.) found at Nisa and other sites on the southern borders of Turkmenistan; 60 the 
first-century B.C. ostraca from Qumis in eastern Iran; 61 the first-century a.d. parchment 
from Awroman in Kurdistan; inscriptions on the coins of the Arsacid kings of the first half 
of the first century a.d.; the bilingual inscription of Seleukia (150-15 1); 62 the inscription 
of Ardavan V found in Susa (215); some third-century documents discovered in Dura- 
Europos; the inscriptions at Kal-i Jangal, near Birjand in eastern Khurasan (first half of the 
third century); 63 inscriptions of the Early Sasanian kings in Parthian; and the writings of 
the Manichaeans. The Parthian script and language began to be abandoned from the fourth 
century. With the exception of Manichaean literature, which appears in its own particular 
script (see above), all the above-mentioned items are in the Parthian script, which is an 
adaptation from the Aramaic with several variants. Harmatta has provided a table showing 
all of these. 64 

Part Two 

(A L. Khromov) 

Borrowings and influences 

The fact of Persia's extensive literary borrowings from India in the Sasanian period has 
been established for some time. A substantial part of those borrowings subsequently passed 
from Pahlavi to Arabic literature and thence to the West. 65 Iranian interest in Indian philos- 
ophy and science during the Sasanian period is demonstrated by translations into Middle 

60 Diakonoff and Livshits, 1976-79. 

61 Bivar, 1970; 1972; 1981. 

62 Morano, 1990. 

63 Henning, 1958, p. 40. 

64 Harmatta, 1958, p. 175. 

65 Ol'denburg, 1907, pp. 49-50. 



ISBN 978-92-3-103211-0 Contents Borrowings and influences 

Persian of Indian works on mathematics, astronomy and medicine, and of belles-lettres and 
didactic texts (see above). 

During the Sasanian period much effort was devoted to the translation of Indian works 
on astrology that predicted the future and described natural phenomena, such as the flight 
of birds and the cries of animals, in terms of good or bad omens. Some Arab sources 
report that the Sasanian court was visited by sages from India who advised on future events 
according to the position of the planets, the signs of the zodiac and the configuration of 
the stars. Their predictions were written down, conveyed to the king and subsequently 
preserved in a secret depository. Tradition has it that over 100 doctors, including several 
from India, were employed at the court of Khusrau I (531-579). It may thus be assumed 
that Indian medical works were translated in Sasanian Iran. 66 

The bulk of scientific literature in the Sasanian period, however, was translated from 
Syriac and Greek. To understand why this was so it is necessary to explain the position of 
Christianity in the Sasanian Empire. In the western part of the empire (in Mesopotamia), 
Zoroastrianism coexisted with Christianity and Judaism and with the worship of ancient 
Babylonian, Greek and Syrian gods, but Christianity gradually became the main rival of 
the Zoroastrian clergy and the Iranian administration which it controlled. The new religion 
spread throughout the vast territory of Sasanian Iran, and by the seventh century Christian 
communities were to be found in nearly every province of both western and eastern Iran. 

At the beginning of the third century there were still ethnic and linguistic differences 
between Zoroastrians and Christians in Iran: most Zoroastrians were Iranians while the 
majority of Christians belonged to other ethnic groups (Syrian, Greek, Armenian). Between 
the fifth and the seventh centuries, however, the Christian communities in central and east- 
ern Iran underwent a process of Tranization' and the Iranian ethnic component became 
dominant. Relations between Zoroastrians and Christians can be summed up as follows. 
The Sasanian secular administration showed an extremely pragmatic attitude towards the 
Christians, harsh repression coexisting with the presence of Christians at court and even 
their acceptance as marriage partners. The Zoroastrian clergy, on the other hand, were 
fiercely intolerant of Christians at all periods, and were restrained only by the political 
and economic interests of the Sasanian rulers and their circle. Zoroastrians did not accept 
the alien Christian teachings, although they did not express open hostility towards the 
adherents of other faiths. 67 Each or all of these factors operated at some period during 
the Sasanian Empire and determined the extent to which literature in translation was able 
to circulate in Iranian society. 

66 Inostrantsev, 1907, pp. 73-7. 

67 Kolesnikov, 1988, pp. 23-5. 



ISBN 978-92-3-103211-0 Contents Science and philosophy 

The Syriac-Nestorian literature of the pre-Islamic period developed to a considerable 
extent alongside the literature of the official Sasanian religion from roughly the fifth cen- 
tury onwards. Its authors were mainly Iranians. 68 The principal Nestorian schools in 
Sasanian Iran were in Nisibis (Mesopotamia), Ctesiphon, the Sasanian capital, and Gunde- 
shapur (the Syriac Bet Lapat) in Khuzistan. The school in Nisibis was established following 
the closure in 489 of the Nestorian 'Persian school' in Edessa (founded in the fourth cen- 
tury by the Syrian, Ephraim), after which all the teachers and pupils moved to the territory 
of the Sasanian Empire. The Nisibis school, which was a theological academy enjoying 
special privileges, produced a number of Nestorian scholars who made an important con- 
tribution to the history of ideas in the East. 69 Its rector, Rabban, was also professor of 
biblical exegesis. His deputy, whose responsibilities included instruction in Bible reading 
and the liturgy, also taught philosophy. The school had a tutor and a secretary. When Nisi- 
bis was taken by the Arabs in the seventh century, the school had some 800 students. 70 

The school of Gundeshapur was known less as a theological college than for its hos- 
pital (bimdristdn) and the medical academy attached to it. The hospital was founded in 
the reign of Shapur I (241-271) but reached its maximum expansion in the sixth century 
under the patronage of Khusrau I. When the emperor Justinian closed the Academy in 
Athens in 529, the staff emigrated to Gundeshapur. The Nestorians who had been banished 
from Byzantium became energetic propagators of Greek education: it was to them that 
the school in Gundeshapur owed its world reputation. Students not only acquired a the- 
oretical training based on the works of Galen but also participated in the medical work 
of the hospital. 71 The hospital survived until the beginning of the c Abbasid period in 
the eighth century. Many famous doctors from the school of Gundeshapur subsequently 
worked in Baghdad. 72 

Science and philosophy 

Under the Sasanians, medicine was based on the spirit of the Zoroastrian religion but it 
was also influenced by Greek medicine. Great importance was attached to healing by 
the power of words, using magic formulae taken from sacred books. According to the 
Denkard, a doctor was required to have a good knowledge of anatomy, organic func- 
tions and the properties of medicine and also to be attentive to his patients. Since he was 

68 Baumstark, 1968, p. 100. 

69 Pigulevskaya, 1979, p. 67. 

70 Fuck, 1981, p. 276. 

71 Ibid. 

72 Siassi, 1963, p. 370; Endress, 1987, pp. 407-8. 



ISBN 978-92-3-103211-0 Contents Science and philosophy 

expected to visit the sick as many times a day as necessary, he had to be well fed, and 
provided with a good horse and a comfortable place to stay in the centre of the town. 
He should not be grasping: a good doctor was considered to be one who practised for 
religious reasons. There were, however, rules governing payment for medical care. The 
payment depended on the property and social position of the patient and on whether the 
whole body or only specific organs were treated. The doctor was required to provide the 
sick with regular and painstaking treatment; his refusal to examine a sick person was con- 
sidered a crime. Doctors possessed a type of licence authorizing them to engage in med- 
ical practice. When an Iranian doctor was available, it was considered wrong to consult 
one of foreign origin; despite this, Sasanian kings often preferred to use Greek doctors or 
Syrian Christians. 

The medical community had its own hierarchy. The first distinction was drawn between 
doctors who ministered to the spirit and doctors who ministered to the body. The for- 
mer belonged to the same caste as the priests. Above the drustbadh (the state's chief doc- 
tor) stood the mobadan mobad (the state's leading religious dignitary). The autobiography 
of Burzoe (a famous physician during the reign of Khusrau I) was included by Ibn al- 
Muqaffa c in his preface to the Arabic translation of the Kalilag u Dimnag. Burzoe gives a 
description of the medical literature of his day which testifies to the influence exerted by 
Indian medicine on that of the Sasanian period. The third book of the Denkard contains a 
medical treatise compiled from sources going back to the Sasanian era. 73 The tolerant atti- 
tude of the Sasanian court towards religions under Khusrau I, and the benevolent attitude 
of this ruler towards Christian scholars, did much to encourage the translation of works 
written in other languages. The Middle Persian translation of the Old Testament Psalms, 
fragments of which have been found in East Turkestan, was produced during this period. 
The basis for the translation was the Syriac text. 74 

It was Khusrau I, above all, who encouraged the development of the sciences in Iran 
and the use of Greek and Indian sources. Khusrau was extremely interested in philosophy 
and, in particular, the ideas of Aristotle and Plato. This interest was partly responsible for 
the appearance of a work by Paul of Persia, the Prolegomena to Philosophy and Logic, 
and the Commentaries on Aristotle's De interpretatione and Analytica priore which the 
same author dedicated to Khusrau I Anushirvan, 'Shahanshah, Benefactor of the People'. 75 
At the court of Khusrau II (590-628), Ava of Kashgar was renowned for his knowledge 

73 Christensen, 1944, pp. 414-20. 

74 Ibid., p. 422. 

75 Christensen, 1944, pp. 422-3; Pigulevskaya, 1979, p. 147; Endress, 1987, p. 408. 



ISBN 978-92-3-103211-0 Contents Science and philosophy 

of the Greek, Persian, Syriac and Hebrew languages and as a specialist in medicine and 
astronomy. 76 

Greek sources on astrology were reworked on the basis of Indian theories; parameters 
and astronomical calculations were taken over from Indian works. Works on astronomy and 
astrology by Greek authors were translated into Middle Persian as early as the reign of Sha- 
pur I. The Persian historian Ma c na, the Catholicos of Seleukia under Yazdgird I (399-420), 
translated Greek works into Syriac and then from Syriac into Persian. Catholicos Akakios 
(484- 496), also from Seleukia, translated the Syriac work Discourse on Faith 11 for Kavad 
I (488-531), who adopted a tolerant attitude towards Christians. 

Our knowledge of Middle Persian scientific literature is predominantly based on Arabic 
translations and on information and quotations culled from Arab sources. The surviving 
Arabic versions of Middle Persian texts are based on late adaptations and exhibit the com- 
bination of Greek and Indian components that are typical of astrology in the Sasanian 
period. One example is the collection of tables for use in mathematical astronomy com- 
piled during the reign of Khusrau I and re-edited under Yazdgird III (632-651). The later 
version was used by the Arab scientists MashaTlah and Abu Macshar, who compiled a 
work entitled the Zlg as-Sahriydr [Astrological Tables of Shahriyar], in which the combi- 
nation of Greek and Indian astronomical theories can be clearly traced. In Arabic, the word 
zig became a term denoting a textbook on astronomy or astrology. 78 

A work entitled the Varz-namag, which contained an account of basic agricultural prac- 
tice, was also translated into Middle Persian from Greek. 79 It is assumed that the Middle 
Persian text of the Aln-namag, which has come down to us in the Arabic translation and 
contains information on the military theory of the Sasanians, was composed under the 
influence of an anonymous Greek treatise and the Strategikon of Maurikios of Byzantium 
(c. 600). 80 The Syriac language was thus the link that enabled the Near and Middle East 
to assimilate the achievements of Greek science, which enjoyed a new period of creativity 
on Arab and Persian soil. 81 Literature translated into Middle Persian during the Sasanian 
period played an important part in this process. 

76 Endress, 1987, p. 408. 

77 Krymskiy, 1905, p. 59. 

78 Endress, 1987, pp. 413-14. 

79 Ibid., p. 414. 

80 Inostrantsev, 1909, p. 65; Pigulevskaya, 1946, p. 33. 

81 Pigulevskaya, 1979, p. 31. 



ISBN 978-92-3-10321 1-0 Contents Court chronicles and epic histories 

Court chronicles and epic histories 

The Sasanian period saw the revival of the ancient Iranian tradition of court chronicles, the 
essence of which was to interpret historical events in the light of the king's wishes. The 
content of historical chronicles was restricted to events which the king considered impor- 
tant for himself, his family and the state. They were composed under the king's supervision 
and were intended for his own personal use and that of his heirs. 82 The historical works of 
the Sasanian period focus less on contemporary events than on the past and on predictions 
of the future. Their purpose was to describe and extol the religious and national ideals 
propagated by the Sasanian elite. 

There was no clear dividing line in historical works between fact on the one hand, 
and myth and legend on the other. Such works were not impartial accounts of events but 
included fantasy and emotional declarations. Hyperbole and metaphor were essential ele- 
ments of their style. In the works of the Sasanian period, historical figures were endowed 
with the characteristics of contemporaries. For example, all the pre-Sasanian kings acquire 
Sasanian features and deliver speeches from the throne, as was done in the Sasanian period. 
References are made to sites of the Sasanian period and the historical circum-stances 
described are also contemporary. 83 At the same time, these chronicles were sources of 
ethical guidance for all Iranians. They contained sage reflections and exhortations as well 
as examples of wise decisions by monarchs and their courtiers. 

The Kdrndmag-T ArdasTr-T Pdbagdn [Book of the Deeds of Ardashir Papakan] belongs 
to the epic cycle on the Sasanian kings. Although the text was initially drafted in the sixth 
century, the version which has survived dates from a later period. The main character in 
the Kdrndmag is Ardashir I (226- 241), the founder of the Sasanian Empire. The work 
describes the childhood and youth of Ardashir, his struggle for power and his ascension 
to the throne; it also tells of his son Shapur I and grandson Hormizd I (271-272) and of 
Ardashir's conflict with the Parthian king Ardavan V (c. 213-224). Although the account 
of Ardashir's life given in this work is largely based on legend, it reflects some histori- 
cal events and facts: information is provided about Ardashir's campaigns against nomadic 
tribes, the history and geography of Iran and the social structure and religious conceptions 
of Sasanian society. 

Various versions of the history of Ardashir I are to be found in works by Arab authors 
of the eighth to the tenth century and in Firdausi's Shah-name. These variants, which differ 
from the text of the Kdrndmag, go back to the Late Sasanian collection of histories of 

82 Klima, 1977, pp. 41-3. 

83 Yarshater, 19836, pp. 402-3. 



ISBN 978-92-3-10321 1-0 Contents Court chronicles and epic histories 

the Iranian kings, the Khwaday-ndmag, in which genuine historical facts are closely inter 
woven with legend. The royal chronicles were drawn on for descriptions of the events of 
the Late Sasanian period. 84 

The Khwaday-ndmag and other historical works were translated into Arabic by Ibn 
al-Muqaffa c (757). The Fihrist of Ibn al-Nadim contains a lengthy list of Arabic transla- 
tions of works of the Sasanian period. It is thanks to such translations and adaptations by 
authors of the Islamic period, and to references to them in the works of medieval Arabic- 
and Persian-speaking writers, that we have some idea of their content and nature. One of 
these works was the Gdh-namag which, according to the Tanbih of al-Mas c udi, contained 
a description of the boundaries of the Sasanian Empire. Another work, entitled the Aln- 
ndmag [Book of Rules], describes the customs, morals and behaviour prescribed for kings, 
aristocrats and other high echelons of society. According to the information given in the 
Arabic and Persian sources, these works were of a rhetorical nature and were composed in 
the Iranian tradition. There were also a number of works on the administrative system and 
on individual kings and national heroes. 

The Karnamag-l Anosirvan [Book of the Deeds of Anushirvan], a series of fragments 
presenting the thoughts and utterances of Khusrau I, is reproduced in full in the Arabic 
work the Tajdrib al-umam by Maskawaih. It describes the hostile intentions of political 
and religious sectarians towards the king and his efforts to preserve the traditional dis- 
tinctions in society which divided warriors from peasants. The text contains information 
about the king's relations with other peoples, especially the Turkic Khazars to whom he 
extended his protection and dispatched Mazdakite missionaries. It also describes advances 
in legislation and cultural borrowings from Greece and India in spite of the religious 

The Testament of Ardashir Papakan, which has been preserved in the Istanbul manu- 
script, 85 belongs to the Late Sasanian period. It explains many aspects of royal power 
and touches on various questions involving the relationship between royal power and 
religion. 86 

The literature of the Sasanian period, particularly the historical chronicles and the 
andarz (wisdom) literature, devotes much space to the image of the ideal king, which 
was first conceived under the Achaemenids. The basis for the idealization of royal power 
is the notion that the Persian kings are called on to embody the 'national Iranian idea', the 
essence of which is that all world history is to be seen as a struggle between two primary 

84 Chukanova, 1987, pp. 9, 11. 

85 Grignaschi, 1967, pp. 1-2. 

86 Menasce, de, 1983, p. 1,183. 



ISBN 978-92-3-10321 1-0 Contents Court chronicles and epic histories 

principles: good and evil, light and darkness. In this struggle, it is the king's role to be the 
supreme commander of the army of Ohrmazd (i.e. the good principle) in the war against 
the forces of Ahriman (i.e. the evil principle). 87 This also determines the king's position in 
society and his personal qualities: he is the focus of absolute power and controls history; 
he surpasses all other men in physical strength, looks, intelligence and eloquence; he is the 
lawgiver and creator of order. All kings are not only great military leaders but also talented 
politicians and thinkers. They are heroes in battle, the fount of prosperity, builders of cities, 
creators of all that is useful in the administrative structure of the state and, most important, 
energetic defenders of the Zoroastrian faith. 

According to the mobad Kartir, the 'ideal king' should be religious above all: he must 
be completely subservient to his religious mentor and act and think in accordance with the 
dogmas of the Zoroastrian religion. 88 The king's main occupations were the administra- 
tion of justice, in consultation with the mobadan mobad, courtiers of high rank and wise 
counsellors; the resolution of problems relating to peace and war, and the appointment of 
military commanders; the enactment of measures to ensure the country's prosperity; and 
the settlement of questions relating to hunting, banquets and weddings. 89 As the people's 
spiritual leader, the king was responsible not only for the country's administration but also 
for the regulation of its ethical and social life. He was answerable for his own conduct and 
for that of the government. 

The above qualities were reflected in the king's messages, testaments and pronounce- 
ments as well as in the messages of his ministers. 90 His first responsibility was to deliver 
a speech from the throne to his courtiers. Rendering thanks to God, he set out his plans, 
assuring the people of his desire to rule justly and asking for their support. The assembled 
courtiers approved the king's speech and his intentions and assured him of the people's 
obedience. The ceremony was conducted with great pomp. The king sent messages to 
regional governors, informing them of his ascension to the throne. 

Among the Sasanians, Ardashir I, Shapur II (309-379) and Khusrau I were endowed 
with the traits of the ideal king. There are many written accounts in medieval Arab and 
Persian sources describing Khusrau I as a just and generous king. A typical story (based, 
in Christensen's view, 91 on a reliable source) is included in the work by Nizam al-Mulk 
entitled the Siydsat-name 91 In the story, Anushirvan the Just (i.e. Khusrau I) delivers one 

87 Knauth and Nadjmabadi, 1975, p. 202. 

88 Lukonin, 1969, p. 110. 

89 Yarshater, 19836, pp. 406-7. 

90 Yarshater, 19836, p. 399. 

91 Christensen, 1944, p. 369. 

92 Siydsat-name, 1949, pp. 34-42. 



ISBN 978-92-3-103211-0 Contents Religious life 

of his speeches from the throne in which, addressing his courtiers, he instructs them to deal 
generously with 'the people of God', to lighten the burden of the people, not to offend the 
weak, to respect the wise and to be attentive to good people. The king threatens to punish 
those who disregard these commandments. 93 

Religious life 


In Sasanian Iran, religion played a central role in the life of society, regulating the entire 
spectrum of social and political life and official standards of behaviour. The Zoroastrian 
church was unified under the Sasanians and acquired considerable political power. Its priv- 
ileged position greatly assisted the Sasanians' rise to power. The Sasanians and the Zoroas- 
trian church were united by the idea of centralization - the power of the shahanshah and 
of the mobaddn mobad were two expressions of the same view. The inscriptions of Kartir 
testify to his power and influence. 94 In spite of a degree of rivalry and occasionally strained 
relations, church and state shared the same world view and the same aims. The state usually 
supported the church and often helped to eradicate heresy. The church in turn supported 
the structure of the state, the privileges of the elite, the divine right of the shahanshah and 
the belief in complete obedience to him. Young people were nurtured in the ideals of the 
monarchy, which was underpinned by and oriented towards the church. 95 Ardashir I issued 
the following admonition to his son Shapur I when the latter was preparing to ascend the 

O my son, Religion and the State are sisters. They cannot survive without each other. Religion 
is the buttress of the State and the State is its protector. And whatever is deprived of support 
crashes down and whatever is not defended is lost. 96 

Zoroastrian temples in which the sacred flame was tended were to be found throughout the 
empire. One of the principal temples was located in Atropatene (in Azerbaijan), and in it 
was preserved the flame of the king and warriors. Another principal temple was situated in 
Fars (the flame of the priests) and a third in Khurasan (the flame of the farmers). A major 
role in the establishment of a unified state and church was played by the mobad Kartir, 
whose career began under Shapur I and reached its zenith under Shapur's successors in the 
years 273-293 when he was mobaddn mobad and spiritual director of the king. 97 

93 Ibid., pp. 35-6. 

94 Yarshater, 1983a, p. xxxiv. 

95 Yarshater, 1983a, p. xviii. 

96 Klima, 1957, pp. 40-1. 

97 Istoriya Irana, 1977, p. 110. 



ISBN 978-92-3-103211-0 Contents Religious life 


In the course of the third century a new religion, Manichaeism, appeared in the Sasanian 
Empire (see Chapter 17). Its founder was Mani (216-276), who was descended from 
a notable of Iranian stock on his mother's side. Manichaeism, which expanded ancient 
Iranian conceptions of the eternal struggle between the kingdoms of light and darkness, 
also incorporated elements of Christianity, Gnosticism and Buddhism. According to the 
teaching of the Manichaeans, the world was a chaotic mixture of dark and light elements. 
In order to free themselves from the power of the devil, human beings had to cleanse 
themselves of evil and had therefore to escape from the power of the material principle. 98 
Manichaeism, which spread widely throughout Mesopotamia, Iran, the Roman Empire and 
Central Asia, gradually came to resemble a kind of 'Protestantism', and to be an ideolog- 
ical weapon against evil in the world." The Manichaeans (ZandTks) formed the princi- 
pal heretical group in the Sasanian period and were frequently persecuted by the official 
church. Mani died a martyr's death on 20 March 276. 10 ° 

Mani chose Middle Persian as the vehicle for the dissemination of his religion in the 
Sasanian Empire. Appearing at the court of Shapur I, he presented the king with a book 
entitled the Shdbuhragdn [Book of Shapur], which contains a concise exposition of Mani- 
chaean doctrine. The work was originally written by Mani in Aramaic (his native tongue) 
and then translated into Middle Persian. Although all of Mani's canonical works were 
subsequently translated into Middle Persian, most of these translations have been lost. 

One of Mani's most important works is the Evangelion [Gospel], which in the Middle 
Persian version has this Greek title. It comprises 22 chapters, each of which begins with 
a different letter of the Aramaic alphabet. A fragment of the introduction to this work has 
been preserved in which Mani refers to himself as Yisho Aryaman (an apostle). The use of 
such an Iranian term indicates that the Manichaeans were concerned to make their teaching 
more accessible to Zoroastrians. 101 

The Middle Persian versions of Mani's other canonical works - the Niydni zindagdn 
[Treasure of Life], the Pragmateia and the Rdzdn [Mysteries] - have been lost. Fragments 
of the Book of the Giants exist in three languages: Middle Persian, Parthian and Sogdian. 
A collection entitled the Epistles has been preserved: it consists of letters written by Mani 
to various preachers of his community. In order to propagate their faith, the Manichaeans 
borrowed several stories and fables from India and China which subsequently reached the 

98 Ibid 

99 Lukonin, 1969, p. 71. 

100 Klima, 1957, p. 41. 

101 Boyce, 19836, p. 1198. 



ISBN 978-92-3-103211-0 Contents Religious life 

West. They thus played an active role as intermediaries in the transmission to Europe of 
Eastern fables. 102 

Most Iranian- language versions of Manichaean works have been preserved in Sog- 
dian and only a small proportion in Middle Persian and Parthian. There are collections 
of prayers and two long psalms composed by Mani. The Parthian text of one of these 
psalms, entitled the Vuzurgdn AfrTwdn [The Blessing of the Great], has survived in good 
condition. A few Sogdian and Middle Persian fragments of this psalm have also been 
preserved. The second psalm, which is called the Qsudagdn AfrTwdn [The Blessing of 
the Consecration] in Parthian, is similar to the first in structure and content. 103 Several 
Manichaean manuscripts possess good illustrations and it is believed that there was a text 
entitled the Ardhang which contained commentaries on pictures. Iranian-language frag- 
ments of a text composed in the style of canonical tradition have also been preserved: 
it is known by the Greek title Kephalaia. The Middle Persian version of one text is a 
translation of a Christian apocryphal text entitled The Shepherd of Hermes, which was 
borrowed by the Manichaeans and used as an allegory of the life of man. There are also 
texts recounting Mani's last meeting with Bahram I (272-275) 104 and Mani's death 105 - 
both by eyewitnesses and dating from 274-276 or thereabouts. Middle Persian, Parthian 
and Sogdian fragments of texts have also been preserved which present the story of the 
Manichaean church and Mani's life, listing his works and recounting the early missions of 
his followers. 106 

Some Manichaean prose works contain the rules of the Manichaean faith, fragments of 
homilies and exhortations in the form of questions and answers (a type of oral 
didactic literature encountered in Zoroastrian works). Certain works (mostly in Middle 
Persian) have a cosmogonic content. The fragment of a creed, works on astronomy and 
calendars are also extant. 107 A large number of Manichaean texts are hymns, poetic works 
performed to music. Two long cycles of Manichaean hymns in Middle Persian entitled the 
Gowisn Tg Griw Zindag and the Gowisn Tg Griw Rosn are dedicated to the embodiment 
of the Manichaean deity: this is referred to as 'the living being itself and, according to 
Manichaean beliefs, embodied all elements of light scattered throughout the world. 

Middle Persian Manichaean literature is notable for the eclecticism of its content, struc- 
ture and style. Manichaean hymns, which were influenced by the ancient Iranian tradition 

102 Henning, 1945. 

103 Boyce, 19836, p. 1200. 

104 Henning, 1942. 

105 Boyce, 19836, p. 1201. 

106 Boyce, 19836, pp. 1201-2; Sundermann, 1971a; 19716; 1974. 

107 Henning, 1947; Boyce, 19836, p. 1202. 



ISBN 978-92-3-103211-0 Contents Religion and the law 

(especially the Parthian texts), may be considered as something akin to poetry. 108 Art, 
poetry and language developed in all the areas to which Manichaeism spread, particularly 
where it became the state religion, if only for a limited period of time. But Manichaeism 
was a blatant heresy from the Zoroastrian viewpoint. It was also perceived as such by 
the other religions which, although alien to each other, united in the struggle against the 
Manichaeans. 109 

Religion and the law 

The history of religious struggles under the Sasanians throws light on the underlying causes 
of many events involving politics and ideology. The persecution of reformers and heretics 
- at times, intense; at others, less harsh - is an indication of the state religion's desire to 
maintain its supremacy. For Iranian Zoroastrians there was an intimate connection between 
law and religion. Evidence of this is provided by the MddTgdn-T hazdr dddestdn [Book of 
a Thousand Judicial Decisions], which was written in Middle Persian in c. 620 and has 
survived in a single manuscript. Its author, Farrukhmard, from the town of Gur in the 
province of Ardashir-Khvarreh, was a contemporary of Khusrau II. As codified law did 
not exist in Sasanian Iran, the book cannot be considered a legal code. It is one of the 
collections that were compiled as manuals for the administration of justice. In addition 
to general information about the rights of specific state departments and officials, such 
manuals contained passages from official edicts and decrees. In the specialist literature, 
the Mddlgdn has become known as the 'Sasanian Legal Code'. It has been established that 
some of the sources of this and similar codes of laws of the Sasanian period reflect statutes 
recorded in the legal nasks (precepts) of the Avesta and in the commentaries on those nasks, 
the so-called chastaqs. The Code contains references to certain edicts of Kavad I (488-53 1) 
and Khusrau I; it also contains a remarkable instruction concerning the appropriation by 
the royal treasury of the property of the Manichaeans and their followers. In compiling the 
Code, Farrukhmard made use of legal records from the town of Gur as well as private legal 

It is possible to reconstitute practically the entire system of Iranian law on the basis of 
the mass of information contained in the Code. Its content also contributes to an under- 
standing of other Pahlavi texts including the Denkard, the Dddestdn-T denig, the andarz 
and rivdyat and a number of Middle Persian and Parthian epigraphic texts. The text of the 
Code is also important for the study of the Ishobokht [Code of Laws], which has survived 



Boyce, 19836, pp. 1203-4. 
Bausani, 1965, pp. 50-1. 



ISBN 978-92-3-103211-0 Contents Religion and the law 

in a Syriac translation and contains legal norms of the Christian communities in Sasanian 
Iran; and for the study of the Babylonian Talmud, which describes the law of the Jewish 
communities of the Sasanian Empire. 110 

There were sizeable Jewish communities in the towns of Mesopotamia and Iran: that in 
Iran was self-governing and was generally not persecuted by the Sasanians. 111 The sage 
Samuel was a trusted adviser of Shapur I, and the mother of Shapur II lent her support 
to the Jewish rabbis. The only outbreak of persecution directed against the Jews occurred 
during the reign of Yazdgird II (438-457), particularly in Isfahan which was the centre of 
the Jewish community, but it was of short duration. Nevertheless some Jews subsequently 
emigrated to Arabia and India. 112 

110 Perikhanyan, 1973, pp. xiii-xxiv. 
1 ' ' Istoriya Irana, 1977, p. 111. 
112 Klima, 1957, p. 44. 



ISBN 978-92-3-103211-0 Contents State organization and administration 


A. H. Dani and B. A. Litvinsky 


State organization and administration 107 

Economy, society and trade 112 

Religious life 113 

Cities, architecture, art and crafts 115 

Languages and scripts 121 

In the early centuries of the Christian era the names of two great empires stand out boldly 
in the history of Central Asia. The first was Kushanshahr, named after the Great Kushan 
emperors, who held sway from the Amu Darya (Oxus) valley to the Indus and at times as 
far as the Ganges. Here flourished the traditions of the Kushans, who had brought together 
the political, economic, social and religious currents of the time from the countries with 
which they had dealings (see Chapter 7). The second great empire (which rose to challenge 
Kushan power) was Eranshahr, which expanded both westward and eastward under the new 
Sasanian dynasty. Its eastern advance shook Kushan power to its foundation. 

State organization and administration 

According to Cassius Dio (LXXX, 4) and Herodian (VI, 2.2), Ardashir I (226- 241), who 
waged many wars, intended to reconquer those lands which had originally belonged to the 
Persians. He defeated the Parthian kings and conquered Mesopotamia - an event which 
led to his wars with the Romans. It is more difficult to judge his conquests in the east. 
According to the inscription of Shapur I (241-271) at Naqsh-i Rustam, 'under the rule of 

See Map 2. 



ISBN 978-92-3-103211-0 Contents State organization and administration 

shahanshah [king of kings] Ardashir' were Satarop, king of Abrenak (i.e. Abarshahr, or 
Nishapur, in Khurasan), as well as the kings of Merv, of the Sakas and of Kerman, all 
of whom were called Ardashir'. 1 But al-Tabari's information is different. He describes 
the conquest of Seistan (modern Sistan), Abarshahr, Merv, Balkh and Khwarizm 'up to 
the farthermost borders of countries of Khurasan'. Then he writes that envoys of the 
Kushans, of Turan and of Makran came to Ardashir and offered their submission. 2 Sha- 
pur's dominions in the east undoubtedly reached Merv and Seistan. To study the prob- 
lem, Harmatta has used Greek, Latin and Armenian sources for the conquests of Ardashir 
I in the east against the background of all the wars that he waged; he has also taken 
into consideration the chronology of the Roman-Sasanian wars. His conclusions support 
al-Tabari's information. Ardashir's military activity in the west lasted from 224 
to 232. 3 

Harmatta 4 has rightly argued that after the conquest by Ardashir I, the western part 
of the former Kushan Empire became a vassal kingdom under the Sasanians, and subse- 
quently a province governed by the Sasanian prince governors. These prince-governors 
issued coins as Kushanshahs (kings of the Kushans). At a later stage, the king of Kabul 
formed a marriage alliance with Hormizd II (303-309). 

During the rule of Shapur I, Ardashir's son and heir, the power of the Sasanians in the 
east increased. This is demonstrated by the titles of Shapur I and the inscription on the 
Ka c be of Zoroaster, according to which: 

the state of Shapur I, 'shahanshah of Iran and non-Iran', included Varkan [modern Gurgan], 
Merv, Harev [Herat], the whole of Abarshahr, Kerman, Segistan [Sistan], Turan [near Kalat 
in Baluchistan], Makuran [Makran], Paradan [near Quetta], Hindustan [Sind], Kusansahr up 
to Puskabur [Peshawar] and up to the boundaries of Kas [Kashghar or Kesh], Sughd and Sas 
[Tashkent]' (SKZ, 2). 5 

1 The inscription text (Frye, 1984, p. 272). Earlier editions: Honigmann and Maricq, 1953; Sprengling, 
1953. The first author pointed out the difference in the titles of Ardashir I and Shapur I in 1979. Noldeke 
has expressed his opinion that information from other sources concerning Ardashir's conquests in the east 
(al-Tabari in particular) is much exaggerated (Noldeke, Tabari, 1973, S. 17-18). 

2 Noldeke, Tabari, 1973, S. 17-18; Frye, 1984, p. 295. However, Noldeke and Frye suppose that most of 
the events took place during the rule of Ardashir's successor, Shapur I. 

3 According to Harmatta, 'the main target of Ardashir's eastern campaign was apparently the Kusana 
Empire' . The king made his campaign to the east in 233-235 and the Sasanian invasion of Bactria occurred in 
233 (Harmatta, 1965, pp. 186-94). But Maricq believes that the eastern campaign took place at the beginning 
of Ardashir's I reign (Honigmann and Maricq, 1953, p. 106). We cannot establish the matter with certainty, 
but Harmatta's interpretation seems the most convincing. 

4 Harmatta, 1969, pp. 386-8. 

5 Some scholars (Lukonin, 1969a, pp. 29-30; Zeimal, 1968, pp. 92-100) doubt the reliability of the 
inscription concerning the description of the territory controlled by the Sasanians. In their opinion the enu- 
meration of provinces was only a pretence, and these lands were not really included in the state of Shapur. 



ISBN 978-92-3-103211-0 Contents State organization and administration 

Under Shapur I, the south-eastern provinces were united administratively and the country 
was called 'Sind, Segistan and Turistan as far as the sea shore'. Narseh (293-303), the son 
of Shapur I, was appointed ruler and he received the title of sk'n MLK' (the king of the 
Sakas). The main dynastic line of the Great Kushans from Kanishka I had come to an end, 
but another line of Eastern Kushans continued to rule (see Chapter 7) in Gandhara and 
the Indus valley. On this interpretation, only the northern part of the Kushan Empire came 
under the direct rule of the Sasanian prince-governors. 

A great number of Kushano-Sasanian coins 6 connected with the Sasanian domains in 
the east have been found in Pakistan, Afghanistan and Central Asia. They may be divided 
into three groups. 7 The first group was minted according to the model of the Kushan coins 
of Vasudeva. They are represented by both bronze and gold coins. On the obverse there 
is a figure of a standing king, each with a distinctive crown on his head and a legend 
consisting of the name and the title §ahano §aho (shahanshah) (Fig. 1) of the ruler. On 
the reverse there is a representation of Shiva and his vehicle, Nandi (the bull), with the 
legend, borzaoando iazado (the exalted deity) (Fig. 2). These coins were minted at Balkh 
and perhaps also at other centres (the mint name Bahlo for Balkh occurs on the obverse of 
some issues). 

The second group, also struck in bronze and gold, follow the Sasanian pattern. On the 
obverse they have a portrait of the ruler with his own individual crown (there are at least 
eight types of crown). On the reverse a fire altar is represented. These coins were minted 
in Merv and Herat; the language of the inscription is Middle Persian, written in Pahlavi 

The third group consists of bronze coins that are much thicker and often of irregular 
shape. Their portraits are similar to those of the second group. The language of the inscrip- 
tion is Middle Persian, in Pahlavi script on some issues, in the Bactrian alphabet on others. 
Many of these coins have been found in Pakistan and it has been suggested that they were 
issued and circulated there. However, they have also been found to the north of the Hindu 
Kush and in Tajikistan. 

However, other scholars (Harmatta, 1969, pp. 420-9; Livshits, 1969, p. 56) consider that the inscription 
on the Ka c be of Zoroaster is a reliable historical source and all these lands, including Kushanshahr, were 
included in the Sasanian Empire. 

6 Cunningham, 1893; Bivar, 1956; 1979; Herzfeld, 1930; Lukonin, 1967; 1969a, pp. 39-40; Cribb, 1981; 
Zeimal, 1983, pp. 257-61; Gobi, 1984, pp. 70-86; Trever and Lukonin, 1987, pp. 64-9; Harmatta, 1969, pp. 
385-7, 430; and others have contributed to their study. 

7 Lukonin has suggested that the first group should be called 'Sasanian-Kushan', the second 'Kushano- 
Sasanian' and the third 'the Group of Kavadh' (Lukonin, 1967). 



ISBN 978-92-3-103211-0 


State organization and administration 


FIG. 1. Kushano-Sasanian coins (observe). Photo: © Bibliotheque Nationale de France. 

On the basis of the coinage, the following list of the Kushano-Sasanian kings can be 
established: 8 

Ardashir I Kushanshah 
Ardashir II Kushanshah 
Peroz I Kushanshah 

8 Some scholars have pointed out that the name of Hormizd on the coins corresponds to that of Hormizd 
II (302-309), shahanshah of Iran. This provides a number of synchronisms while typological analysis has 
made it possible to identify the context of other coins. The concept was developed in detail by Herzfeld, 
1930, and Bivar, 1979. 



ISBN 978-92-3-103211-0 Contents State organization and administration 


FIG. 2. Kushano-Sasanian coins (reverse). Photo: © Bibliotheque Nationale de France. 

Hormizd I Kushanshah 

Peroz II Kushanshah 

Hormizd II Kushanshah 

Varahran I Kushanshah 

Varahran II Kushanshah. 



ISBN 978-92-3-10321 1-0 Contents Economy, society and trade 

However, Lukonin and his followers have advanced a number of serious arguments 
against this sequence and its chronology. His outline agrees with that of Gobi. 9 

Economy, society and trade 

Several excavations undertaken in Kushanshahr provide ample material to reconstruct dif- 
ferent aspects of contemporary life. The economic base may be inferred from the currency. 
Although gold and silver coins are known for several Kushano-Sasanian prince-governors, 
it is the copper coinage that was widely current to meet the local demands of the popula- 
tion. Herzfeld has shown that gold continued to circulate in international markets, as it did 
in the time of the Great Kushans. 10 This must be due to the continuity of trade that followed 
the Silk Route as in the earlier period. The only change was that whereas the Kushans had 
earlier controlled this trade, it was now the Sasanians who, in spite of their wars with the 
Romans, were in charge of the principal flow of goods. Movements of the nomads in the 
east may have hampered the movement of caravans but the excavated materials suggest 
that goods were still freely exchanged from one country to another. 

The best evidence for such international trade contacts comes from Begram, Taxila and 
Dalverzin-tepe, where imported goods from several countries have been found together. 
As Ghirshman rightly remarks: 

There was a transit trade in spices from China and Arabia, and nard and pepper were exported 
from India. International trade encouraged the growth of colonies of merchants, particularly 
Jews and Syrians, who established themselves as far afield as India, Turkestan, Brittany and 
the Black Sea. The exporting houses became more specialized, and confined their dealings to 
cor, cattle, and manufactured goods. 11 

Another indication of the extent of international contacts is provided by the free move- 
ment of missionaries, Christians, Manichaeans and Buddhists, who travelled long distances 
from the eastern Sasanian provinces across Transcaucasia to Xinjiang along the frequented 
routes and promulgated their faiths. 

The social picture of Kushanshahr is extremely complicated because of the movements 
of several nomadic tribes into the territories ruled by the Kushano-Sasanian governors. 

9 According to Lukonin, the conquest of the Kushan lands falls in the last decade of Shapur II (309-379) 
and the investiture relief of Ardashir II (379-383) at Taq-i Bustan represents the victory of the Sasanian 
shahanshah over the last Kushan king. Gobi comes to very similar conclusions but dates the beginning of the 
coinage to an earlier period - 356 - in his major work (1984). Other hypotheses refer the Kushano-Sasanian 
coinage to the fourth century. It is not possible to resolve the problem at present. However, the present author 
(Dani) supports Bivar's system and chronology. 

10 Herzfeld, 1930. 

11 Ghirshman, 1954, pp. 342-3. 



ISBN 978-92-3-103211-0 Contents Religious life 

The picture will not be complete unless we start a little earlier and follow the process of 
migration until the coming of the Hephthalites. 

Religious life 

As the evidence of the coins clearly shows, the Zoroastrian faith, the traditional religion 
preserved in the province of Fars, enjoyed great popularity among the Kushano-Sasanians. 
Cults of Anahita and Ohrmazd ( Ahura Mazda) continued to occupy a prominent place 
in Fars after the Sasanians came to power. In both south-western Iran and the north-east 
there were great sanctuaries served by priests. The influence of Zoroastrianism can be 
inferred from the representation of fire altars on the coins. The most important sanctuary 
of fire worship is that excavated at Surkh Kotal, which although originally built by the 
Kushan emperor Kanishka, was still in good repair during this period. The influence of 
Zoroastrianism can also be seen in Gandhara sculpture where fire worship is depicted; and 
fire temples have been excavated at Kara-tepe. The Zoroastrian religion in Bactria must 
have been reinforced by the expansion of the Sasanian Empire. 

It was in this period that the cult of Mithra flourished. On the Kushano-Sasanian coins, 
we find the Bactrian legend, borzaoando iazado (Middle Persian: bwld'wndy yzty). Ini- 
tially, this title was given to the Indian god Shiva but he soon received the nimbus with 
rays around his head like Mithra. Mithra was apparently regarded as a divine protector of 
the Kushano-Sasanian rulers and gave them their power. 12 

Ghirshman 13 mentions 'an imperial religion', the need for which he sees in the rise of 
Shapur I. The Sasanian emperor 'needed to mobilize all his national forces for the struggle 
with Rome'; hence 'the sympathetic interest shown by Shapur to the teachings of Mani' 
(for Manichaeism, see Chapters 3 and 17). But the real patron of Mani was the young 
Peroz, by whom he was probably introduced to Shapur. Mani was of noble birth and like 
other prophets he claimed to have been sent by God to fulfil the mission of earlier religions. 
'He preached a universal religion' and 'derived his doctrines from the cults of Babylonia 
and Iran and also from Buddhism and Christianity.' 14 

According to Ghirshman, 'In the East, Buddhism [had been] at the height of [its] expan- 
sion' since the time of the Kushans. 'In the West, centres of Christianity had sprung up in 
northern Mesopotamia and Judaism was active in Babylonia.' Thus Zoroastrianism was 
squeezed by all these surrounding religions within its home province of Fars; and when 
Manichaeism challenged its position even here, there was a violent reaction. ' Mani was 

12 Lukonin, 1987, p. 138. 

13 Ghirshman, 1954, pp. 309-10. 

14 Lukonin, 1987, pp. 69-70. 



ISBN 978-92-3-103211-0 Contents Religious life 

tried, condemned and put to death. His followers were persecuted and fled abroad, some to 
the East, where their teachings flourished in Central Asia.' 15 

How far was Zoroastrianism preserved in the Kushano-Sasanian kingdom? There is 
ample evidence to show the widespread popularity of the fire cult and some scholars 
even maintain that the followers of Zoroastrianism persecuted Buddhists. Ohrmazd was 
still worshipped, however, and the dogmas, with their emphasis on a single divinity, were 
upheld. The mobads (high priests) accepted the traditions of south-western Iran, and along 
with them continued the worship of Anahita and Mithra, although they were given a sec- 
ondary place. The old sacred traditions were put in writing in the form of the Avesta in 
the fourth century. Strengthened by its religious organization, 'Zoroastrianism was able to 
drive out Manichaeism and hold Christianity in check on the line of the Euphrates and 
Buddhism on the Helmand.' 16 

Buddhist missionaries, on the other hand, continued to exert their influence through- 
out the whole of Afghanistan and Central Asia. Their famous centres at Ghazni, Kabul, 
Bamiyan, Balkh and Termez maintained continuous contact with the centres in Gandhara, 
and central and East Turkestan. It seems very likely that Buddhism itself was undergoing 
a great change in its practices, ideological concepts and rituals. With the acceptance of the 
image of the Buddha and the expansion of Buddhism and Buddhist monasteries, the educa- 
tional character of the sangha (Buddhist community) had taken on a new shape. The place 
of the Bodhisattvas and several other local deities had greatly increased. But Buddhism 
was not alone in this new expansion. The evidence of coins suggests that the popularity 
of Shiva and Nandi had caught the popular imagination. At the same time, the seated god- 
dess Ardokhsho had begun to be identified either with Lakshmi or with Hariti, the queen 
of the yakshis. Other Hindu deities such as Karttikeya, Baladeva and Vasudeva were also 
prominent and the worship of the sun god remained important. 

Funerary practices such as corpse position and the burial of bones stripped of flesh and 
soft tissue (which clearly refer to Bactrian Zoroastrian rites) provide important evidence for 
the religious beliefs of the local population. Burials took place in special structures (known 
as naus) as well as in ruins of old buildings. There were also burials in graves in the shape 
of pits. These are all a direct continuation of funerary rites of the Kushan period. 17 

15 Ghirshman, 1954, pp. 314-18 

16 Ibid., p. 318. 


Litvinsky and Sedov, 1984, pp. 75-137. 



ISBN 978-92-3-10321 1-0 Contents Cities, architecture, art and crafts 

Cities, architecture, art and crafts 

Begram, Dalverzin-tepe, Kara-tepe and Surkh Kotal were among the principal founda- 
tions of the time of the Great Kushans. While Begram was the summer capital, Kara-tepe 
presents a group of caves, stupas and monasteries and Surkh Kotal preserved its character 
of a fire temple (probably built by Kanishka). As the excavations showed, Begram was 
sacked by Shapur I in c. 244 but the city continued to exist afterwards and to play a role. 
The Buddhist complex of Karatepe was also deserted at some moment when the Kushano- 
Sasanians occupied it, but the discovery of Kushano-Sasanian coins and later construction 
provides evidence of its subsequent continuity as a religious centre. Similarly, Surkh Kotal 
preserved its character. 

Kara-tepe 18 presents a typical model of cultural material that is syncretistic in nature. In 
original concept, the Buddhist caves and monasteries copy the type that is so well known in 
Gandhara. Staviskiy sums up: 'The Kara-Tepe excavations brought to light wall paintings, 
stone and stucco (ganch in local terminology), sculpture, terracottas, and pottery, metal 
ware and coins, inscriptions on pottery and graffiti on the walls of the caves and their 
entrance niches.' The structures constitute a complex of caves, a courtyard and some grand 
buildings. The ensemble is divided into four groups, A, B, C and D. Groups A and B 
originally formed a single complex with three courtyards, placed next to one another on a 
terrace along the eastern slope. In the northern courtyard, the remains of a stupa have come 
to light, while the central and southern courtyards were bordered with porticoes. South of 
the southern courtyard was a Buddhist temple, consisting of a cella surrounded by three 
corridors. The cave temples consisted of a cella and a vaulted corridor that encircled it. The 
walls in most surface and cave premises of groups A and B were painted over the plaster. 

A fire altar stood in an attached niche with a shell-shaped back wall: the existence of a 
fire altar in a Buddhist structure is most striking. It resembles those of the Kushan period 
found in Gyaur-kala in Khwarizm and Samarkand. Staviskiy remarks that 'altars of this 
kind were used by fire worshippers, and in temples of the local Mazdean cults, and also in 
early Buddhist structures'. They survived up to the period of the Arab conquest. Staviskiy 

After the complex B was destroyed or had fallen into abandonment, a fire altar was built in 
the niche which had previously held a Buddhist statue. It was made hastily, out of materials 
that were at hand and symbolized the victory of some other cults (most likely Zoroastrianism) 
over Buddhism. 19 

18 Staviskiy, 1984, pp. 95-135. 

19 Ibid., p. 114. 


ISBN 978-92-3-10321 1-0 Contents Cities, architecture, art and crafts 

Part of the wall-painting shows a figure of the Buddha seated under a tree with three stand- 
ing monks, close in execution to the Indian style. Here the Buddha has a halo around the 
body, probably of local Bactrian origin. Numerous stone and stucco fragments have been 
found which originally formed part of the decoration. Staviskiy concludes: 

Stupas and cave premises (A and B complexes) were derived from India, along with Bud- 
dhism, whereas the planning of the cave temples is alien to the ancient Indian tradition and is 
rather to be traced to the religious architecture of the Middle East, where similar fire temples 
could be found consisting of a cella and the corridors encircling it. As to the temple court- 
yards of Kara-tepe, they go back to the Rhodian type east Hellenistic courtyards that had 
become common in Bactria and the neighbouring Parthia already in third-second centuries 
B.C. 20 

It may be noted that Buddhist temples of the type seen in Kara-tepe, particularly the plan- 
ning and the representation of the Buddha with the nimbus and halo around the whole 
body, became common in East Turkestan and also in Eastern Asia during the post-Kushan 

Many other cities and settlements are known where life developed in the Kushano- 
Sasanian period on the territory of Bactria ( Tokharistan). At the Yavan site in southern 
Tajikistan, for example, a section of a small street was unearthed on the citadel. On each 
side there was a solid area of large house blocks with many rooms. These houses consisted 
of individual, interconnected premises. Walls were erected from pakhsa blocks (pakhsa is 
a local term for blocks of rammed clay mixed with finely chopped straw) and clay bricks. 
The structures probably had two storeys. 21 The structures at the site of Halkajar, on the 
hills of southern Tajikistan, are similar to those at Yavan. A good collection of terracottas 
was made from here. 22 

Dalverzin-tepe, in the Surkhan Darya region (southern Uzbekistan), was a large city 
under the Kushans; but towards the end of the period, the main parts of the city and its 
fortifications were neglected. However, buildings DT- 6 and DT-7 and the temple in the 
northern part of the city continued to function, as did the potter's quarter with its kilns. 
The naus of Dalverzin continued to fill with funerary deposits, although burials were also 
found outside the naus. Kushano-Sasanian copper coins have been discovered at the site. 
Thus it is clear that the city continued to be inhabited in the traditional manner, although 
activity was not as intense as before. 23 

20 Staviskiy, 1984, p. 133. 

21 Zeimal, 1975, pp. 267-9; Litvinsky, 1973, pp. 14-17. 

22 Sedov, 1987, pp. 80-4. 

23 Pugachenkova et al., 1978. 



ISBN 978-92-3-10321 1-0 Contents Cities, architecture, art and crafts 

Zar-tepe, another large city in the Surkhan Darya region, is of considerable interest. Its 
heyday was in the Kushan period, but subsequently its fortifications fell into disrepair. In 
the centre of the city stood a palace-type complex with two halls: one was four-columned 
(13x9.3 m); the other was twelve-columned (17.6x9.2 m), with an aiwan (hall) in front. 
During the last period, brick pedestal-supports for a fire altar were constructed between 
the columns. The halls adjoined other rooms where fragments of painted clay sculptures 
have been found. Zar-tepe had a Buddhist shrine, a large structure decorated with clay 
sculptures of which one head is preserved. Buddhism was not the only religion of the 
urban population, however. 

A number of residential districts have been excavated. Blocks of adjoining houses were 
discovered on both sides of the street, each block consisting of 8-1 1 interconnected build- 
ings. Every block was separated from its neighbour by a narrow street at right angles to the 
main road. In one of the house blocks a shrine was also found with, in its centre, a fire altar 
standing on a square platform. Similar shrines have been found at other sites. Here again, 
Kushano-Sasanian coins were discovered in the upper level. 24 

Kay Kobad Shah, the ancient capital of the Kobadian district, continued to exist under 
the Kushano-Sasanians, its fifth and last period of existence being separated from the previ- 
ous occupation by a sterile layer. 25 Alongside the large cities a great number of villages and 
other settlements existed. In addition, cult buildings, burial monuments and other construc- 
tions have also been excavated. Even in the small Kobadian oasis many Kushano-Sasanian 
archaeological complexes have been recorded, in particular at the sites of Kay Kobad Shah, 
Ak-tepe II, Darakhsha-tepe, Klych-duval, Shodmon-kala and Munchaktepe, where sanctu- 
aries, burials, remains of irrigation works and other structures have been found. 26 

Among cult buildings, mention should be made of the monastery of Ushtur-Mulla, near 
the Amu Darya river. The monastery consisted of 26 buildings which surrounded a square 
courtyard. In the middle of the north side was a shrine with a H -S haped circumbulatory 
corridor. There were small residential cells for individual monks and a large hall for sangha 
(congregational) meetings. The shrine was decorated with paintings and alabaster sculp- 
tures. Outside the complex was a stupa, the base of which was faced with stone reliefs. 27 

Another site of this period is Chaqalaq-tepe, in northern Afghanistan, which had an oval 
plan and was surrounded by two circles of city walls made of pakhsa. Its two lowest layers 
were dated by coins to the Kushan period. The 307 coins in the upper levels included 72 

24 Masson, 1976; Zavyalov, 1979. 

25 Dyakonov, 1953, pp. 276, 289. 

26 Sedov, 1987, pp. 11-48. 

27 Zeimal, 1987. 



ISBN 978-92-3-10321 1-0 Contents Cities, architecture, art and crafts 

imitations of coins of Vasudeva and a coin of the Sasanian Kushanshah, Varahran II. 28 
Village sites have also been excavated, such as Durman-tepe, where two small hoards of 
Kushano-Sasanian coins were discovered. 29 So, during the Kushano-Sasanian period life 
continued in the cities and villages to the north (that is to say, in 'right-bank' Bactria) and 
south of the Amu Darya. 

The number of sites in both northern and southern Bactria with settlements of the 
Kushano-Sasanian period is extremely large; 30 there were also some in central Afghanistan, 
particularly in the Kabul region. Two Buddhist monasteries have been discovered at Tepe 
Maranjan (Kabul), in one of which a hoard of 326 silver Sasanian coins and 12 gold 
Kushano-Sasanian coins was found. 31 

Some of the architectural features of the period have been mentioned above. There were 
established categories of architectural structure: palaces, common residential quarters and 
various types of cult buildings, including Buddhist monasteries both in caves and on the 
ground. There were also buildings serving technical functions such as water-reservoirs 
and workshops. Monumental structures and ordinary everyday buildings existed side by 
side. They differed greatly in the quality of their construction. Pakhsa, especially for wall 
foundations, and square mud-bricks were the most common materials; baked bricks and 
stone were used more rarely, mainly for paving and for the pedestals of column-bases. 
Roofs were usually flat, and in small buildings the beams rested across the tops of the walls. 
In large buildings, however, columns were also used. Vaults and domes were known. For 
the transition from the square ground plan to the circular dome, squinches were adopted. 
All kinds of passages and niches had arches. The wall surface was covered with one or 
more layers of plaster. The interiors both of monumental secular buildings and of cult 
buildings were decorated with wall-paintings and sculpture. 

There were two trends in the art of the time: Buddhist art developed from traditions 
going back to the art of Gandhara with local features, whereas non-Buddhist art dis- 
played a complex fusion of local and Sasanian traditions. This latter feature is well seen 
in the sculptures and wall-paintings from Dilberjin. 32 Especially interesting are the wall- 
paintings from structure 12 of the northeast cult complex. The largest fragment represents 
a figure on a yellow throne, sitting frontally with knees apart. This pose is typical of repre- 
sentations of Sasanian kings and gods enthroned. The arms also are kept apart and there is 
an object, possibly a mirror, in one hand. The left hand is resting upon a shield. The figure 

28 Mizuno (ed.), 1970. 

29 Mizuno (ed.), 1968. 

30 Sedov, 1987, pp. 11-48, 78-95; Ball and Gardin, 1982, especially p. 483, Map 67. 

31 Curiel, 1953; Hackin, 1953; Fussman and Le Berre, 1976, pp. 95-9. 

32 Kruglikova, 1976, pp. 96-100, Figs. 56-8. 


ISBN 978-92-3-10321 1-0 Contents Cities, architecture, art and crafts 

is dressed in a loose shirt and there is a belt around the waist. A cloak is thrown over the 
shoulders and there is a golden helmet on the head. Black plaits hang beneath the helmet 
and the head is surrounded by a yellow nimbus. (The presence of the nimbus indicates 
that the figure represents a deity.) Two yellow horn-shaped objects project from behind the 
shoulders. On each side there are two standing personages turning towards the deity and a 
number of smaller figures. The outermost personage of the right-hand group is only half 
as large as the other standing figures. The contours are traced with a thin black line. The 
painter understood the application of light and shade, and although frontal representation 
is a dominant feature of this style, profile figures are slightly inclined to the front. 

This painting is a fusion of several traditions. The form of the helmet reflects that of 
Antiquity. There are strong connections with the subsequent paintings in Tokharistan at 
Balalyk-tepe and Kala-i Kafirnigan. This provides clear evidence that a local Bactrian art 
existed and was developed in Bactria during Kushano-Sasanian times, although it absorbed 
other traditions and was powerfully influenced by official Sasanian art. 

Fine specimens of metal-working of the Kushano-Sasanian period are known. In Sasa- 
nian toreutics, Lukonin has singled out an 'East Iranian School' which he ascribes to mas- 
ters at the court of the Kushanshahs. Among works of this school he lists five silver dishes, 
upon each of which an equestrian hunt is represented. The dish in the Hermitage Museum is 
remarkable (Fig. 3). It depicts a rider, hunting wild boar. One beast has already been struck 
by the rider's sword, the other is emerging from the undergrowth. On the rider's head there 
is a crown with spiral horns. The crown resembles in general terms that of Varahran II 
Kushanshah. The surface of the images is gilded and there is a Sogdian inscription of the 
fifth-sixth centuries on the dish, providing an upper limit for the date. Lukonin concludes 
that Kushano-Sasanian metal-work follows the general trends of Sasanian art, particularly 
in the choice of subjects such as heroic hunting scenes, but the composition differs in exer- 
cising more freedom. 33 The orientation of Kushano-Sasanian metal-work towards Sasanian 
art can be explained by its court character. 

The minor arts are mostly represented by terracotta figurines. Only occasional examples 
follow the traditions of Sasanian art, especially those of Sasanian portraiture. Bactrian fea- 
tures emerge in the foreground in this rather popular art. Craft industry, especially pottery, 
was highly developed. There are examples of table-services, storage jars and kitchen ware. 
Some types were developed from those used in Kushan times, but new types continued to 
emerge. The ornament of vessels changed, and became more concentrated towards the top 
of the vessel. 

33 Trever and Lukonin, 1987, pp. 61-73, 108, PL 14 (7), 15 (7). 



ISBN 978-92-3-103211-0 


Cities, architecture, art and crafts 

FIG. 3. Silver dish. Photo: © Terebenin (Hermitage, St. Petersburg). 

Everyday utensils, ritual objects, mirrors of several types, buckles and ornaments were 
made of bronze. Tools and weapons were made of iron. Pins, combs and similar items 
were made of bone and there were beads, and vessels made of glass. A large quantity 
of jewellery was inlaid with a variety of precious stones. Armourers produced composite 
bows, arrowheads, daggers, swords, lances and plates of armour. The shape of certain types 
of weapon changed, especially the triangular arrowheads, which developed a groove behind 
the point. 34 

34 For details, see Sedov, 1987. 



ISBN 978-92-3-103211-0 Contents Languages and scripts 

Languages and scripts 

It is only possible to judge the language of the Bactrians in ancient times by personal 
names. The situation changes, however, in the first centuries a.d. The discovery of the 
inscriptions at Surkh Kotal, and then other inscriptions, especially in Central Asia, has 
allowed scholars 35 to determine the peculiarities of the Bactrian language. It belongs to 
the group of East Iranian languages, more precisely to the north-eastern group. There is a 
marked similarity with the modern Munji language, which is probably a continuation of 
one of the Bactrian dialects. Bactrian represents the Middle Iranian stage. 

Written Bactrian used an adaptation of the Greek alphabet, with an additional 25th 
letter, san, which reproduced the sound s. The Greek writing reproduced the Bactrian 
phonetics only approximately, although certain methods of Greek orthography were trans- 
ferred into Bactrian and Greek scribes apparently took part in creating the Bactrian writing- 
system. Initially its letters were of lapidary or 'monumental' style, but later a cursive form 
was developed. In spite of the Sasanian conquest, Bactrian scribes continued to work in 
Balkh and other centres. By the third century the transition from a lapidary to a cursive 
style was completed. Round- shaped letters and the presence of ligatures are characteristics 
of the cursive style. Kushano-Sasanian coin legends, as well as the inscriptions on wall- 
paintings from Dilberjin, are of this type. Some of them can be referred to the third-fourth 
centuries. The inscriptions include explanations of the content of the paintings. 36 

In the state of the Kushanshahs, apart from high officials and military personnel, there 
were inevitably natives of Iran, including both scribes and marginally literate persons. It 
is quite natural that they wrote their Middle Persian in Pahlavi script 37 (see Chapter 3), 
so that some issues of the Kushano- Sasanian coinage (see above) have Middle Persian 
inscriptions in Pahlavi script. At the Buddhist complex of Kara-tepe ( Termez), Middle Per- 
sian inscriptions in Pahlavi script (written in the monumental style without ligatures) were 
found on the cave-walls. Specific palaeographic features suggest that these inscriptions go 
back to the fourth-fifth centuries rather than to the third. One inscription originally had a 
date, but it is now destroyed. Another bore the one-line record, zyk dpyr (Zik, the scribe). 38 
There are also inscriptions in Gandhan Prakrit and Buddhist Hybrid Sanskrit. 

35 Maricq, 1958; 1960; Henning, 1958; 1960; Gershevitch, 1967; Humbach, 1966-67; Harmatta, 1964; 
1965; 1969; Livshits, 1969; 1975; 1976; Fussman, 1974; Steblin-Kamenskiy, 1981; Lazard et al., 1984. 

36 Livshits, 1976. 

37 Saleman, 1900; Nyberg, 1964; 1974; Rastorgueva, 1966; MacKenzie, 1971; Gignoux, 1972; Rastorgueva 
and Molchanova, 1981. 

38 Lukonin, 1969b. 



ISBN 978-92-3-103211-0 Contents Languages and scripts 

Most inscriptions of the Kushan and Kushano-Sasanian periods come from Termez, in 
particular from the Buddhist monasteries of Kara-tepe and Fayaz-tepe. They are written in 
the KharosthT and BrahmT scripts, mostly on potsherds. In Kara-tepe alone over 100 pot- 
sherds with inscriptions were discovered, besides 7 wall graffiti. 39 The use of both scripts 
was current at the same time. 

The BrahmT alphabet has indications for the vowels. The consonants with the inher- 
ent a are divided into twenty-five mutes, falling into five classes, four semi-vowels, three 
sibilants, one aspirate, one pure nasal and three voiceless spirants. The combination of 
vowels and consonants is represented by ligatures. According to Dani: 'This alphabetic 
system is maintained in India with minor additions or omissions down to the present day, 
though it is not phonetically suited to the various provincial languages in India.' The evolu- 
tion of BrahmT continued for many centuries. Considerable changes took place during the 
first-fourth centuries, among which the introduction of new forms of signs should be men- 
tioned. BrahmT is read from left to right, the same direction as that of European alphabets. 40 

As Dani explains: 'The whole system of KharosthT follows the pattern of Aramaic. The 
resemblance is not so much in the identity of forms, though a few letters are the same, 
but in the way in which these forms are produced.' 41 The palaeography of Central Asian 
inscriptions in KharosthT has its own peculiarities. There are two types of inscription, one 
following the tradition of stone inscriptions, with typical straight lines, the other resembling 
the manuscripts of Central Asia. 42 

As Vorobyova-Desyatovskaya observes: 

It is also important to point out the differences between the wording of the KharosthT and 
Brahmi inscriptions of Kara-tepe. Most of the KharosthT inscriptions are based on a tradi- 
tional formula indicating the donor, the gift, the recipient, almost resembling an incantation 
formula. However, the BrahmT inscriptions, even though a dedication, mention a sarigha, 
vihdra [monastery] or school. They are focused on the donor and his attributes. Others again 
serve as indications of the individual use of a vessel. 43 

The inscriptions from Fayaz-tepe are similar: 44 

The comparison of Indian inscriptions on the territory of Central Asia with those from India 
and Afghanistan shows that as a rule they were worded according to a general pattern. The 
palaeography of the Kara-tepe and Fayaz-tepe inscriptions confirms the presence of a stan- 
dard scribal tradition for the entire Kushan territory. 45 

39 Vertogradova, 1983, p. 3. 

40 Dani, 1963, pp. 75-104. 

41 Ibid., pp. 265-7. 

42 Vertogradova, 1982, pp. 150-1. 

43 Vorobyova-Desyatovskaya, 1983, p. 24. 

44 Vertogradova, 1984, p. 167. 

45 Vorobyova-Desyatovskaya, 1983, pp. 51-2. 



ISBN 978-92-3-103211-0 Contents Origin and rise of the Kidarites 



E. V. Zeimal 


Origin and rise of the Kidarites 123 

Early history of the Kidarites 124 

Foundation of the Kidarite state 125 

Kidara's conquest of Gandhara and Kashmir 126 

The decline of the Kidarites 127 

The new conquest of Bactria by the Sasanians (467) 129 

Economy, society and polity 130 

Monetary system and trade 132 

Life-style, culture and ideology 136 

Origin and rise of the Kidarites 

The fourth century witnessed the appearance in Central Asia of a new wave of nomadic 
tribes known by a variety of names and recorded in numerous sources. By the Latin authors 
they were called Chionites, and by the Greek authors Kidarite Huns or 'Huns who are 
Kidarites'; in Indian chronicles they were known as Huna, in Armenian literature as both 
Honk' and Kushans, and in the Chinese annals as the Ta Yiieh-chih, or Lesser Yiieh-chih 
(i.e. the same name that denoted their forerunners in those lands, who founded the Kushan 
Empire). Their ruler was called Kidara or Chi-to-lo (Ancient Chinese, kjie-td-ld). The use 

* See Map 3. 



ISBN 978-92-3-103211-0 Contents Early history of the Kidarites 

of different names sometimes makes it difficult to evaluate the information provided by the 
written sources 1 and to reconstruct a general historical picture. An important prerequisite 
is to make a clear distinction between the Kidarites and another tribal group known as the 
Hephthalites (Hua, I-ta, Hep'tal, Tetal, Hephtal, Abdel and Hayatila in the sources; see also 
Chapter 6). 2 

The terms Huns and Chionites seem to reflect the general ethnic appellation of this 
people, 3 whereas Kidarites should be understood as a dynastic designation derived from 
the name of their king, Kidara. Kushan (widely used in the Armenian sources to designate 
the tribes and state of the Huns) and Ta Yiieh-chih (used in the Chinese sources) refer to the 
country where they established their kingdom, and may reflect their claims to be successors 
to the Kushan kings. The primary basis for identifying the Huns or Chionites with the 
Kidarites is the fact that they are called Kidarite Huns (or 'Huns who are Kidarites') by the 
fifth-century Byzantine author and historian Priscus. 4 None of the other information we 
possess (including numismatic data) contradicts this identification. 

The earliest report on these peoples dates from c. 350 when, according to Ammianus 
Marcellinus (XVI, 9.4), the Chionites (i.e. the Kidarites) fought in Syria as allies of the 
Sasanian king, Shapur II (309-379), at the siege of Amida (the modern Diyarbekir). They 
were led by Grumbates, 'the new king of the Chionites, a middle-aged man, his face already 
deeply lined, possessing an outstanding intellect and famous for the multitude of his victo- 
ries. With him was his son, a fine young man, who fell in the battle' ( Ammianus Marcelli- 
nus, XVIII, 6.20; XIX, 1.7-11). The Chionites (i.e. the Kidarites) assisted the Sasanians 
because of an alliance they had concluded with Shapur II (Ammianus Marcellinus, XVII, 
5.1), who at that time was driven to war against the enemy on the eastern borders of his 

Early history of the Kidarites 

Nothing is known about the history of the Kidarites before the second half of the fourth 
century. It has been suggested that they conquered K'ang-chii and Sogdiana in c. 300, 

1 Of the existing works, the most significant on Kidarite numismatics are Martin, 1937; Curiel, 1953; 
Gobi, 1967; and Enoki, 1969; 1970. See also Ter-Mkrtichyan, 1979. 

2 Trever, 1954. 

3 Neither here nor elsewhere does this term have any ethnolinguistic significance, since we know prac- 
tically nothing definite about the language(s) of these peoples. Just as in the twelfth-thirteenth century the 
term 'Mongol' referred not only to those who spoke the Mongol language proper but also many other ethno- 
linguistic groups forming part of that society, so too could Huns and Chionites have apparently 'incorporated' 
ethnic groups speaking a variety of languages. 

4 The accounts of the Kidarites in this source have been subjected to repeated scrutiny. They were exam- 
ined and compared with the data from Chinese sources by Enoki, 1969. 



ISBN 978-92-3-103211-0 Contents Foundation of the Kidarite state 

but the literary sources have not yet been corroborated by the archaeological evidence. 
Attempts have been made 5 to link the appearance of the Kidarites in Central Asia with the 
Karshi steppe in southern Sogdiana and to date this event to c. 420, but this has not been 
supported by further archaeological investigations in the region. 6 

The only corroboration of the presence of the Kidarites in Sogdiana is provided by early 
Sogdian coins (see also pages 128 et seq.) with the image of an archer on the reverse and 
the word kydr (Kidara) in the obverse legend. 7 These coins were minted in Samarkand 
from the first to the fifth century. But out of some 2,000 such coins, only 7 bear the name 
of Kidara, indicating that Kidarite rule was short-lived. The chronology of early Sogdian 
coinage of the archer type helps to date the coins with the name of Kidara, which cannot 
be earlier than the middle of the fourth century - thus the conquest of Sogdiana by the 
Kidarites cannot have occurred prior to this time. 

Foundation of the Kidarite state 

The alliance between the Kidarites and Shapur II did not apparently last long. The nature 
of the relations between the Kidarites and Sasanian Iran is indicated by the Kidarite issues 
of silver drachms, copying coins of the Sasanian kings. 8 Martin 9 takes the Kidarite coins of 
the Sasanian type as evidence of the existence of a Kidarite state as early as the last quarter 
of the fourth century and dates these coins as close as possible to the time when their 
Sasanian prototypes were issued. He suggests that the Sasanians recognized the Kidarite 
state, while the Kidarites themselves accepted Sasanian suzerainty, and puts other events 
- the Kidarite invasion of the Kushano-Sasanian kingdom, the Sasanian invasion under 
Shapur II and the ousting of the Kushano-Sasanian princes - in the same historical context, 
c. 350. He suggests that the Sasanians recognized the Kidarite state, while the Kidarites 
themselves accepted Sasanian suzerainty. 

Enoki 10 has shown that the establishment of the kingdom of Kidara took place some- 
what later, and Gobi 11 has dated the prototypes of the Kidarite drachms to the period of 
Shapur II and Shapur IE (383-388). 

5 Kabanov, 1953; 1977. 

6 Isamiddinov and Suleimanov, 1984. 

7 Zeimal, 1978, p. 208; 19836, p. 251. 

8 Cunningham, 1895, was the first to make this observation. 

9 Martin, 1937. 

10 Enoki 1969; 1970. 

11 Gobi, 1967. 



ISBN 978-92-3-10321 1-0 Contents Kidara's conquest of Gandhara and Kashmir 

Kidara's conquest of Gandhara and Kashmir 

The most detailed account of Kidara's reign is provided by the Chinese chronicle, the Pei- 
shih [Annals of the Wei Dynasty], written in 643 and covering events between 386 and 
581. 12 The original nucleus of the Kidarite state was the territory of Tokharistan (now 
northern Afghanistan and southern Uzbekistan and Tajikistan), which was previously part 
of the Kushan Empire and subsequently of the Kushano-Sasanians. The capital of the 
Kidarites, the city of Ying-chien-shih, 13 was probably located at the ancient capital of 
Bactria, 14 near Balkh. The lands of the Kidarites were known in Armenian sources as 
'Kushan lands'. 

The Pei-shih relates that Kidara, having mustered his troops, crossed the mountains and 
subjected Gandhara to his rule, as well as four other territories to the north of it. 15 Thus, 
during Kidara's reign, the Kidarite kingdom occupied vast territories to the north and south 
of the Hindu Kush. According to another passage from the Pei-shih, referring to the Lesser 
Yiieh-chih, the principal city of the Kidarites south of the Hindu Kush was situated near 
presentday Peshawar and called (in its Chinese transcription) Fu-lou-sha (Ancient Chinese, 
pyou-lou-sa, which probably represents Purushapura). Its ruler was Kidara's son, whose 
name is not mentioned. 

Historians have found it difficult to determine the exact period of Kidara's reign, one 
reason being that, from the second half of the third century to the fifth century, news reach- 
ing China about events in the Western Regions was generally sporadic and patchy. Li Yen- 
nien, the author of the Pei-shih, writes that 'from the time of the Yan-wei (386-550/557) 
and Chin (265-480), the dynasties of the "Western Territories" swallowed each other up 
and it is not possible to obtain a clear idea of events that took place there at that time'. 
A painstaking textual analysis enabled Enoki (and Matsuda before him) to establish that 
information about Kidara in the Pei-shih was based on the report of Tung Wan sent to the 
West in 437. 16 From this we can infer that, although the Chinese sources do not provide 

12 Information about Kidara in another Chinese dynastic history, the mid-sixth century Weishu [History 
of the Wei], was borrowed entirely (with a few divergences in the transcription of place names) from the 
Pei-shih and has no significance of its own. 

13 In the Weishu, the Kidarites are said to have transferred their capital to another city called Liu-chien- 
shih. However, Enoki has clearly shown that both cities are in fact one and the same, because the accounts 
regarding the Kidarites' transfer of their capital were misunderstood by the author of the Weishu, who tried 
to explain why the capital of the Ta Yiieh-chih was called by a different name at the time of the Han and the 
Northern Wei (Enoki, 1969, pp. 8-10). 

14 The 'city of Balaam', referred to as the capital of the 'Kidarite Huns' by Priscus of Panium, evidently 
corresponds to Bactria-Balkh. For geographical details of events relating to the struggle between the Sasani- 
ans and the Kidarites and then the Hephthalites in the fifth century, see Marshak, 1971. 

15 Zurcher, 1968, p. 373; Enoki, 1969, p. 8. 

16 Enoki, 1969, pp. 8-9. 



ISBN 978-92-3-103211-0 Contents The decline of the Kidarites 

any dates in connection with Kidara, the Pei-shih describes the situation as it existed in 
c. 437. On the other hand, Kidara's rise to power, the founding of his state and the annexa- 
tion of the territories to the south of the Hindu Kush (including Gandhara) should be dated 
to an earlier period, that is to say, some time between 390 and 430, but probably before 

The Kidarites' advance to the south-east apparently continued even in the middle of the 
fifth century. This is indirectly proved by Indian inscriptions depicting the events which 
befell the Gupta king, Kumaragupta I (413-455), when a considerable portion of central 
and western Panjab was under Kidarite rule. Thus, it was in the first half of the fifth century 
that the greatest territorial expansion of the Kidarite state occurred. 

The decline of the Kidarites 

Indian sources, which refer to all nomadic conquerors (Kidarites and Hephthalites alike) 
as Hunas, cast little light on the final stage of the Kidarite state, during which it came 
into conflict with the Guptas. Having captured Gandhara, the Kidarites apparently tried to 
build on their success and extend their territories eastwards into India. The only evidence 
of the war between the Kidarites and the Guptas is the mention of Huna invaders in Indian 
inscriptions referring to the reign of Skandagupta (455-467). 

The first encounter between the two rival powers apparently took place in the reign 
of Skandagupta's father and predecessor, Kumaragupta I. On the evidence of the Bhitari 
pillar inscription, towards the end of that king's reign the Gupta state was on the verge 
of extinction. In this critical situation, Kumaragupta I put his son and heir in command 
of the army with the task of restoring the country's power. 17 Before Skandagupta and his 
army had 'established his lineage that had been made to totter' (line 14), he suffered many 
hardships. According to the inscription, there were times when he had to 'spend a night 
sleeping on the bare earth' (line 10). 

It seems that the enemies who threatened the very existence of the Gupta state included 
not only their feudatories, the Pushyamitras, but also the Kidarites. In the Junagadh inscri- 
ption, 18 which dates from c. 457 and also refers to the reign of Skandagupta, it is obviously 
the Kidarites (or the Hephthalites; see pages 141-2) who are referred to under the name 
of the Mlecchas. Skandagupta's victories over them were described as the conquest 'of the 
whole world'. It is hard to establish what this claim implies in terms of geographical fact, 
but it appears that Skandagupta's armies repulsed the Kidarite invasion somewhere on the 

17 Fleet, 1888, No. 13, lines 10-14. 

18 Ibid., No. 14. 



ISBN 978-92-3-103211-0 Contents The decline of the Kidarites 

River Sutlej (or perhaps further east). Thus, even after Skandagupta's victories, central and 
western Panjab probably remained in the hands of the invaders, 19 although he had managed 
to stop their advance eastwards. 

Yet another inscription - the Kahaum inscription (460^-6 1) 20 - already describes 
Skandagupta's reign as 'peaceful' and calls him the 'commander of a hundred kings'. On 
some silver coins, Skandagupta is given the honorific title of VTkramdditya earlier bestowed 
on his famous grandfather, Chandragupta II (c. 375-413). Numismatic evidence shows, 
however, that war with the Kidarites and other enemies had exhausted the strength of the 
Guptas: considerably fewer gold coins were minted under Skandagupta than under his pre- 
decessors, and the quality of the metal was poorer. Despite the victories over the Hunas, the 
inscriptions show that the Gupta state had already lost a considerable area of its Western 
Territories by the beginning of Skandagupta's reign. The most striking proof of this is the 
total absence of Gupta coins in the western regions of India and in Pakistan. 21 

The war between the Guptas and the Hunas is reflected in a semilegendary form in other 
Indian sources, in particular in Book XVIII of Somadeva's Kathdsaritsdgara (a collection 
of folk tales in Sanskrit literature in the eleventh century), where these events are described 
from roughly the same viewpoint as in the inscriptions of Skandagupta's reign. Despite the 
triumphant tone of the inscriptions, the might of the Guptas declined after these encounters, 
while the Kidarite state continued to exist in western Panjab. 

It was probably not Skandagupta's victories but a new wave of nomadic invaders from 
the north - this time Hephthalites - that put an end to the Kidarite state in Gandhara and 
Panjab. The surviving sources give no clues as to how power passed from the Kidarites to 
the Hephthalites 22 - whether as the result of a clash between opposing armies or of the 
overthrow of one dynasty by another (in a 'palace revolution'). It is known, however, that 
the name Kidara was kept, although now as an honorific title (meaning 'honoured', 'hero', 
'valiant'), long after the Kidarite state had ceased to exist, 23 just as the original Kidara 
used to style himself on coins Kusana Sahi (king of Kushan) many years after the fall of 
the Empire of the Kushans. 

19 Altekar, 1954, p. xxxiv. 

20 Fleet, 1888, No. 15. 

21 Allan, 1914, p. xlix. 

22 This is probably not the event narrated in the 97th chapter of the Pei-shih, in the section dealing with 
the territory of the Lesser Yiieh-chih, whose capital was the city of Fu-lou-sha or Peshawar: 'Their king was 
originally a son of the Great Yiieh-chih king Chi-to-lo. When Chi-to-lo had moved westward under pressure 
of the Hsiung-nu, he ordered his son to hold this city' (Ziircher, 1968, p. 373). 

23 Harmatta, 1984, pp. 188-9. 



ISBN 978-92-3-10321 1-0 Contents The new conquest of Bactria 

The new conquest of Bactria by the 
Sasanians (467) 

During the second quarter of the fifth century the Kidarite state was again involved in 
serious fighting to the west with the Sasanian kings, who could not accept the loss of 
the territories they had conquered earlier from the Kushan Empire. Yazdgird II (438-457) 
took active measures to restore them to Sasanian rule and several eastern campaigns were 
undertaken during his reign. The struggle continued under Hormizd III (457-459) and also 
under Peroz (459-484). 

Western sources - Armenian, 24 Greek, Syriac and others - provide much information 
about the deeds of Yazdgird II and subsequent Sasanian kings in the east but, as mentioned 
above, give different names for Iran's enemies. Often the 'land of the Kushans' or 'Kushan 
regions' are mentioned, which suggests that the term 'Kushan' may be understood not so 
much in an ethnic sense as to mean the inhabitants of the 'land of the Kushans', that is the 
former territory of the Kushan Empire (or, more narrowly, the Kushano-Sasanians). The 
Hephthalites, who are mentioned from the second half of the fifth century, appear on the 
historical scene here, as in India, much later than the Kidarites and as a tribal group distinct 
from and apparently sometimes hostile to them. 

Western, and especially Armenian, authors had a somewhat vague conception of the 
geography of the areas where the campaigns of Yazdgird II and Peroz took place, and 
this has led to conflicting readings and a confusion that is reflected in the conclusions of 
modern researchers. 25 The theory that these battles took place not only in the territory of 
northern Afghanistan but also near the Caspian Sea 26 has proved implausible. 27 There are 
no grounds for identifying two arenas of military action: all the fighting occurred mainly 
in, or in the immediate vicinity of, Tokharistan (one of Yazdgird II' s battles was at Mervi 
Rud), and the Sasanians' main stronghold in the east was Merv. 

Yazdgird II's first eastern campaign was in 442. By 449 the advantage was on his side, 
and the Kushanshah (king of the Kushans) had been rendered powerless. The Sasanian 
army laid waste territories subject to the Kidarites and took fortresses. During the campaign 
of 450, Taliqan was captured. Nevertheless, the struggle continued, and sometimes the 

24 For a collection of references to the Sasanian wars in the east in Armenian sources, see Ter-Mkrtichyan, 
1979, but bearing in mind their evaluation by Lukonin; 1969a; 1969£>. 

25 For example, there were two Chols (one in the region of Gurgan and one in that of Darband), two 
Taliqans (one in Khurasan and one in Tokharistan), etc. Some names have been found to be distorted. For 
further details, see Marshak, 1971. 

26 Marquart, 1901, pp. 55, 211-12; Mandel'shtam, 1958, p. 72. 

27 Marshak, 1971. 



ISBN 978-92-3-103211-0 Contents Economy, society and polity 

balance was in favour of the Kidarites. Their refusal to pay tribute to Yazdgird, mentioned 
by Priscus, prompted renewed military action by the Sasanians in 456. As a result, the 
Kidarites came close to losing all their territories in Tokharistan. This was prevented only 
by the outbreak of civil war (457-459) between Yazdgird II' s sons - Hormizd III, who 
had succeeded his father on the throne of Iran in 457, and Peroz, who at that time ruled as 
governor in the 'Kushan regions'. The war between the brothers continued for two years 
and ended in victory for Peroz, who owed it to the help of the Hephthalites and handed 
Taliqan over to them. 

When war with Iran broke out again in the 460s, Balkh (Po-ho or b'dkld in the Ancient 
Chinese sources) was in the possession of the Kidarites, if we accept that the town of 
Balaam mentioned by Priscus is in fact Balkh. 28 At that time, the ruler of the Kidarites 
was Kunkhas, whose father (the source does not name him) had earlier refused to continue 
to pay tribute to the Sasanians. Peroz, however, no longer had the strength to continue the 
eastern campaign; in 464, according to Priscus, the envoys of Peroz turned to Byzantium 
for financial support to ward off invasion by the Kidarites but it was refused. 

In an attempt to put an end to the war, Peroz made peace overtures to King Kunkhas, 
offering him his sister's hand in marriage, but sent him a woman of lowly birth instead. The 
deception was soon discovered and Kunkhas decided to seek revenge. He asked Peroz to 
send him experienced Iranian officers to lead his troops. Peroz sent 300 of these 'military 
instructors', but when they arrived Kunkhas ordered a number of them to be killed and sent 
the others back mutilated to Iran, with the message that this was his revenge for Peroz's 
deception. The ensuing war against Kunkhas and the Kidarites ended in 467 with the cap- 
ture of their capital city of 'Balaam'. It appears that the Hephthalites were again Peroz's 
allies, as they had been in his struggle with Hormizd for the throne of Iran. This put a final 
end to Kidarite rule in Tokharistan. After their defeat the Kidarites were probably forced 
to retreat to Gandhara, where, as previously mentioned, the Hephthalites again caught up 
with them at the end of the fifth century. 

Economy, society and polity 

The written sources give no details of the Kidarite conquest of Tokharistan, Gandhara and 
the other regions that came under their rule. However, archaeological data are available 
for the study of the Kidarite state, although only for certain areas and tentatively. In certain 
regions of the right bank of the Amu Darya (northern Tokharistan) many towns were largely 

28 It has been suggested that the cities of Balaam and Bolo refer to the site of Er-kurgan in the Karshi steppe 
of southern Sogdiana (Kabanov, 1953; 1977). 



ISBN 978-92-3-103211-0 Contents Economy, society and polity 

destroyed and whole oases laid waste during the last quarter of the fourth century and the 
beginning of the fifth. It is probable that this was caused by military and political strife 
rather than by social and economic upheavals resulting from the natural development of 
society. The huge town of Shahr-i Nau (40 km west of the modern Dushanbe), which came 
into being in Kushan times under Vima Kadphises or Kanishka I and was surrounded by 
a strong defensive wall that was 7 km long and more than 8 m high, with towers every 25 
m, was abandoned at the turn of the fifth century like many other settlements in the Hissar 
valley. Since the valley was beyond the reach of the Sasanian armies, the destruction and 
laying waste of the area is likely to have been connected with the Kidarite invasion. In 
Kobadian, on the other hand, the abandonment of the Bishkent valley and the Shah oasis 29 
and the destruction of Kay Kobad Shah and a number of other towns and settlements can 
be attributed both to the Kidarite invasion and to the Sasanian occupation. 

The picture is similar in the region of the Surkhan Darya, where the site of Dalverzin- 
tepe and other settlements were laid waste at this same time (in particular, in oases such as 
Band-i Khan on the right-bank tributaries of the Surkhan Darya). Here also, it can only be 
assumed that the destruction of economic life was connected with the Kidarites, but so far 
no positive traces of their presence have been found either in the Surkhan Darya valley or 
in southern Tajikistan. 

The Buddhist religious centre in Old Termez, destroyed probably in the 360s-370s by 
the Sasanians, already lay in ruins; yet the mass burials of the victims of an epidemic, or 
some sort of catastrophe, in the abandoned buildings and caves of these monasteries date 
from the time of the Kidarites. More such examples can be given - the late fourth to the 
fifth century was obviously a violent time in northern Tokharistan. If such devastation and 
massacres were indeed the result of the arrival of the Kidarites (which is as yet only a 
hypothesis), then having passed through these regions 'with fire and the sword', they must 
very soon have moved south, 30 beyond the Amu Darya, where they were victorious over 
the Sasanian Kushanshahs. 

The picture is rather different in the Karshi steppe, where archaeological evidence 
reveals a considerable change in the composition of the population in the fourth century. In 
particular, this can be seen from a sharp increase in modelled ceramics with characteristics 
typical of the Kaunchi-Dzhety-Asar archaeological sites (on the middle reaches of the Syr 
Darya, or Jaxartes). Further investigations should clarify whether this is connected with 

29 Litvinsky and Sedov, 1983. 

30 On the right bank of the Amu Darya a large number of tombs (including tumuli) have now been inves- 
tigated, but the vast majority of them belong to the Tokharo-Yiieh-chih and Kushan periods (Litvinsky and 
Sedov, 1984). So far, the right bank of the Amu Darya has yielded hardly any tumuli that can even tentatively 
be linked to the Kidarites. 



ISBN 978-92-3-10321 1-0 Contents Monetary system and trade 

the arrival of large groups of nomads from the north (who might well have been Kidarites) 
in the Karshi oasis. It is clear, however, that the oasis did not suffer destruction and sub- 
sequent abandonment on the same scale as that noted in the regions of southern Tajikistan 
and the Surkhan Darya. 31 Archaeological investigations in other parts of the old Kidarite 
state will doubtless bring to light other patterns of interrelationships between the invaders 
and the local population. 

Our knowledge of the organization of the Kidarite state and the way the conquered ter- 
ritories were governed is just as fragmentary. It is tempting to draw an analogy with the 
vast state of the Kushans. This is not only because the Kidarites claimed to be the succes- 
sors to the glorious Empire of the Kushans; a no less important factor is that the former 
nomadic invaders came into possession of vast territories inhabited by settled agricultural 
peoples with a culture and traditions dating back many centuries, just as had been the case 
with the Tokharians (or Yueh-chih), who created the Kushan Empire. It seems likely that 
the administrative and government structure created by the Kushans was left largely intact 
under the Kidarites. 

Monetary system and trade 

The Pei-shih (Chapters 7, 13) mentions that the Kidarites, whom it refers to as the Ta 
Yueh-chih (Lesser Yiieh-chih), 'have money made of gold and silver' ? 2 This information 
is confirmed by the evidence of their coins. The first comprehensive attempt to catego- 
rize and interpret Kidarite coins was undertaken by Cunningham. 33 Martin, Ghirshman 
and Curiel 34 subsequently made important contributions in the field, but the most detailed 
study of the Kidarite coinage is that of Gobi 35 Gold, silver and copper Kidarite coins are 
now known (Figs 1-3). There are no grounds for maintaining that the Kidarites had a sep- 
arate monetary system as in the case of Kujula Kadphises' early issues; their coinage was 
characterized by an adaptation to the local issues in each area they conquered. In Sogdiana 
small silver coins were issued (drachms reduced to 0.4-0.3 g). They followed the design of 
early Sogdian coins, with the ruler's head facing right on the obverse and a standing archer 
on the reverse, adding the word kydr (Kidara) written in Sogdian on the obverse. 36 

In Tokharistan gold dinars were issued in the name of Kidara, following the gold coins 
of the Kushano-Sasanians both iconographically and technically (on the obverse, a king, 

31 Isamiddinov and Suleimanov, 1984. 

32 Ziircher, 1968, p. 373. 

33 Cunningham, 1895, pp. 55-73. 

34 Martin, 1937; Ghirshman, 1948; Curiel, 1953. 

35 Gobi, 1967; 1984. 

36 Zeimal, 1978, p. 208, PL III, 11; 1983fo, p. 251, PL 21, 10. 



ISBN 978-92-3-103211-0 


Monetary system and trade 

facing left, standing before an altar; on the reverse, Shiva with his bull, Nandi), which 
in their turn can be traced back to the last coins of the Kushan king, Vasudeva I. 37 The 
Tokharian issues of Kidarite coins bear an inscription in Bactrian (Bago Kidara Vazurka 
KosanoSao), with the title 'the great king of the Kushans'. 

The silver coins of Sasanian type can be attributed to Gandhara and the area around. 
They have the ruler's bust facing right 38 or en face on the obverse; 39 and on the reverse, 
the traditional iconographic type for Sasanian coins - a fire altar between two standing 
figures and copper coins of the same type. 40 

FIG. 1 . Kidarite coins with the inscription ij J"vi^. 

37 Gobi, 1967, Vol. Ill, Table 4 - XIII A, XIII B, XIV; 1984, Tables 67-9, Nos. 735-41. 

38 Gobi, 1967, issues 14, 19-24. 

39 Ibid., issues 11-13, 15-18. 

40 Ibid., issues 25-8. 



ISBN 978-92-3-103211-0 


Monetary system and trade 

FIG. 2. Kidarite coins. 

In their Indian territories the Kidarites also issued gold coins 41 based on the model of 
the Late Kushan dinars with the name of Kanishka (III or, if one follows Gobi, II). 42 The 
Kidarite coins of this group bear the name of Kidara written in Brahml script, together 
with the names of dependent rulers or successors of Kidara, on the obverse. The earliest 
coins are the early Sogdian issues with the name of Kidara (not earlier than the mid-fourth 
century). As Gobi has convincingly shown, the Kidarite coins of Sasanian type follow the 
drachms of Shapur II and Shapur III, minted on former Kushan territory, 43 and should 
therefore be dated to the closing decades of the fourth century and the very beginning of 
the fifth. 44 The gold Kidarite coins issued in Tokharistan and in India can be dated more 

41 Gobi, 1984, Tables 46, 47, Nos. 612-15. 

42 Ibid., Table 33, Nos. 538-53. 

43 Gobi, 1967, issues 1-10. 

44 There are no grounds for dating these Kidarite coins of Sasanian model to the time of Shapur II and 
Shapur III and concluding that the Kidarites were originally allies of the Sasanian shahs in their war against 



ISBN 978-92-3-103211-0 


Monetary system and trade 

FIG. 3. Kidarite coins. 

fully. Their issue probably began only in the fifth century (perhaps in its early years) and 
came to an end during the second half of that century. 

As well as gold and silver coins, copper anepigraphic coins, minted on the model of 
the latest copper coins of the Kushan kings, Huvishka, Vasudeva I and Kanishka III (or II), 
were widespread throughout the entire territory of the Kidarite state. They were obviously 
used as small change, as can be seen from the quantities in which they were minted. 45 
These copper coins show the same characteristics as the gold and silver Kidarite coins 
mentioned above - a deliberate adaptation to the existing currency in the conquered terri- 
tories, reproducing (with varying degrees of divergence from the model) the coinage that 

the kingdom of Kushan. Enoki (1970, pp. 34-5) put forward serious arguments against such a supposition: 
Kidara himself had nothing in common with Shapur II, although he used his coins as a model for his own. 

45 For the many discoveries of 'copies' of coins of Vasudeva I and Kanishka III (or II) on the right bank 
of the Amu Darya, see Zeimal, 1983a, pp. 23 1^4, 241-56. Vast numbers of them have also been found on 
the territories of Afghanistan, Pakistan and the north-west of India (Wilson, 1841; Cunningham, 1895; Gobi, 
1976;Cribb, 1981). 



ISBN 978-92-3-103211-0 Contents Life-style, culture and ideology 

was customary in a particular market. As was the case during the reign of Kujula Kad- 
phises, the Kidarites were clearly not yet aware of the political significance of coining 
money. This probably explains, first, why they followed alien iconographic models and, 
second, why certain gold and silver issues - and almost all the copper coins - had no name 
on them (with the exception of the copper coins of the Sasanian model mentioned above 
and issued in the name of Kidarite satraps). 

Thus, the Kidarite monetary system did not disturb the economic life of the regions that 
came under their rule but, on the contrary, created favourable conditions for maintaining 
the established traditions in local trade. The numerous discoveries of imported articles in 
strata of the Kidarite period are proof of the existence of a flourishing international trade 
network and wide trading links between the various regions of the Kidarite state. 

Life-style, culture and ideology 

In the words of the Pei-shih, the Kidarites 'move around following their herds of cattle; 
they also [in this respect] resemble the Hsiung-nu'. 46 On the other hand, it is known that 
there were Kidarite capitals both in Gandhara and in Tokharistan, and thus that they lived 
in towns. This is only an apparent contradiction, as there have been many cases in history 
when nomads (or recently settled nomads), after establishing their rule over large groups 
of states, have, while wholly or partially preserving their traditional life-style, successfully 
adapted to the culture and life- style of the subordinate peoples. 

It would therefore be more accurate to think of the Kidarite state not as a unified society 
but as one with a clear distinction between the conquerors - the ruling group - and their 
subject peoples, the latter preserving their own traditions. It appears that the most impor- 
tant elements in the overall organization and development of the state were the clan and 
tribal organizations traditional to all nomadic peoples; these were inevitably reflected in the 
administrative structure of the state and in the organization of its army - the main support 
of the ruling dynasty. It should be stressed, however, that in such conglomerate states, the 
rulers quickly assimilated the main achievements of the conquered peoples' cultures. The 
Kidarites were no exception, although the written sources mention only the unsuccessful 
attempt of the Kidarite ruler Kunkhas to marry into the Sasanian dynasty (see above). 

The coins make it possible to follow this process of adaptation by the Kidarites in 
much greater detail. We do not know what language the Kidarites spoke, but the coinages 

46 Ziircher, 1968, pp. 373-4. 



ISBN 978-92-3-103211-0 Contents Life-style, culture and ideology 

they issued show inscriptions in Sogdian, 47 Bactrian, 48 Middle Persian 49 and Brahmi. 50 
Kidarite coins display a wide range of iconographic borrowings, reflecting the world of 
Sogdian artistic culture, and the official art of the Sasanians and the art of post-Kushan 
India. Yet it is hard to judge to what degree all these foreign elements penetrated the culture 
of the Kidarites themselves, and how deeply they were assimilated. It should be remem- 
bered that the elements of alien cultures reflected by the coins were the direct result of 
the Kidarites' adaptation to the conditions of the regions conquered by them, and of their 
intention that the coinage they issued should be familiar to the local market. Moreover, 
just as there is no basis for assuming (on the evidence of the coins) that the Kidarites had a 
mastery of all the languages and scripts used in inscriptions on their coins, the iconography 
of Kidarite coins cannot be regarded as a reflection of their artistic tastes. We have hardly 
any knowledge of a specifically Kidarite art that is directly linked to the rulers of the state; 
it is possible to speak only of works of art of the Kidarite period, created by craftsmen and 
artists in the countries conquered by them but following standards and traditions that had 
no direct connection with the conquerors. 

It appears that the Kidarites' beliefs had not yet developed into a rigid religious system, 
which must have encouraged (or at least not hindered) their receptiveness to the religious 
ideology they encountered in the lands they subdued - a local variety of Zoroastrianism 
(Mazdaism) in Tokharistan, various expressions of Buddhism and Hinduism in the terri- 
tory of Gandhara and also, probably, the official Sasanian doctrine. There are no grounds 
for assuming that the Kidarites were Buddhists simply because the Chinese sources report 
the existence of famed Buddhist shrines in the Kidarite capital in Gandhara. At the same 
time, there is nothing to indicate active resistance to any of these religions on the part of 
the Kidarites. During the Kidarite period the Buddhist religious centre in Old Termez lay 
in ruins, like many other Buddhist sites in Tokharistan, but its destruction (as already men- 
tioned) seems to be linked to the religious intolerance of the Sasanians. There are no traces 
of restoration work during the Kidarite period; the partial restoration of the monastery took 
place later, apparently during the second half of the sixth century. 

The key to understanding the ideology of the Kidarite rulers probably lies in their ten- 
dency to consider themselves the heirs of the Kushan kings (many expressions of which 
have been mentioned above). Indeed, this is how they were seen by the neighbouring 
peoples. It is for future investigations (especially in the field of archaeology) to show how 
profoundly and consistently the Kushan heritage was assimilated by the Kidarites. 

47 Zeimal, 1978, p. 208, PL III, II; Zeimal, 1983ft, p. 251, PI. 21, 10. 

48 Gobi, 1984, Tables 67-9, Nos. 735-41. 

49 Gobi, 1967, issue 15. 

50 Ibid., issues 11-28; 1984, Tables 46, 47, Nos. 612-15. 



ISBN 978-92-3-103211-0 Contents Origins 


B. A. Litvinsky 


Origins 138 

Political and military history 141 

Conquests in Gandhara and northern India 144 

The Hephthalites of Central Asia 146 

Social structure and administration 147 

Religious life and polyandry 150 

Language and scripts 151 

Towns 152 

Architecture 153 

Art and crafts 154 


From the mid-fifth to the mid-sixth century Central Asia was ruled by the Hephthalite 
tribes. There are many gaps in our knowledge of the origin of the Hephthalites and the for- 
mation of their state, the first difficulty being that they are given different names in the var- 
ious sources. In Chinese sources the name of the dynasty is I-ta (a variant of I-tien, ancient 
*iep-t'ien) and their king bears the name Yen-tai-i-li-t'o (ancient *Yeptalitha). 1 In Syriac 
sources they are called eptalit, afield; in Greek-language sources, A/38ekai, E<f)Qa\iiar, 
in Armenian sources, hep't'al; in Middle Persian sources, eftal, and also hydn; in Arabic 

* See Map 3. 

1 Enoki, 1959, p. 7. 



ISBN 978-92-3-103211-0 Contents Origins 

sources, haital; and in New Persian sources, hetal? Another name for them is Chinese 
Hua. According to Bal c ami, the etymology of the word 'Hephthalites' is as follows: 'in the 
language of Bukhara', it means 'strong man'. 3 In Khotanese Saka a similar word exists, 
meaning 'brave, valiant'. 

The legends on Hephthalite coins are in the Bactrian script. They feature a Bactrian title, 
XOAAHO, for the ruler together with another Bactrian title, sao. One coin bears the title 
bogo, meaning 'lord' or 'ruler'. The names of Hephthalite rulers given in Firdausi's Shah- 
name are Iranian. Gem inscriptions and other evidence 4 show that the official language 
of the ruling upper class of the Hephthalites in their Tokharistan territories was an East 
Iranian language. 

Chinese sources do not agree on the origin of the Hephthalites. Some hold that they 
originated in Ch'e-shih, that is, from Turfan; others consider them to be 'descendants of 
K'ang-chii' in southern Kazakstan; still others postulate that they were descended from 
the Great Yiieh-chih. The Chinese writer Wei Chieh, who personally conversed with some 
Hephthalites, dejectedly observed: 

However, the information has come from remote countries, and foreign languages are subject 
to corruption and misunderstanding. Moreover, it concerns matter of very ancient time. So 
we do not know what is certain. [In this way] it is impossible to decide [the origin of the 
Hephtalites]. 5 

Information about the physical appearance and language of the Hephthalites also lacks 
precision. For example, Procopius of Caesarea (I, 3) writes: 

Although the Hephthalites are a Hunnish people and are so called, they do not mix and asso- 
ciate with those Huns whom we know, for they do not share any frontier region with them 
and do not live close to them. . . They are not nomadic like the other Hunnish peoples, but 
have long since settled on fertile land. . . They alone of the Huns are white-skinned and are 
not ugly. They do not have the same way of life and do not live such bestial lives as the other 
Huns, but are ruled by one king and possess a legal state structure, observing justice among 
themselves and with their neighbours in no lesser measure than the Byzantines and Persians. 

With regard to language (see also pages 148-9), the Chinese chronicle the Peishih 
reports that 'Their language differs from that of the Juan-juan, Kao-ch'e and various Hu' 
and the account in the Wei shu is similar. The reference to the Hu language testifies to the 
fact that the language of the Hephthalites was distinct from that of those Iranian-speaking 
people of Central Asia who were called Hu by the Chinese. 

2 Altheim, 1959, Vol. 1, pp. 41-3. 

3 Bal c ami, 1869, p. 128. 

4 Maenchen-Helfen, 1959, pp. 227-31; Livshits, 1969, pp. 64-75. 

5 Enoki, 1959, p. 7. 



ISBN 978-92-3-103211-0 



FIG. 1. Afrasiab. Wall-painting. (Photo: © Vladimir Terebenin) 

In the seventh century, after the destruction of the Hephthalite state, Tokharistan was 
visited by a Chinese pilgrim, Hsiian-tsang, who wrote of the Hephthalite population: 

[Their] language and letters differ somewhat from those of other countries. The number of 
radical letters is twenty-five; by combining these they express all objects around them. Their 
writing is across the page, and they read left to right. Their literary records have increased 
gradually, and exceed those of [the people of] Su-le or Sogdiana. 6 

Ibid., p. 39. 



ISBN 978-92-3-103211-0 Contents Political and military history 

This is a clear reference to a Greek-based script of Bactrian origin used in southern Central 
Asia and Afghanistan up to the eighth century. 

In Middle Persian, Byzantine and Indian sources we find the designation 'Red' and 
'White' Huns. This may reflect a division among the Hephthalites or a distinction between 
Hephthalites and Turks. 7 This is also reflected in the mural art: for example, some of the 
envoys in the scene of the Hephthalite embassy in the Afrasiab palace are ruddy-faced, 
while others are pale (Fig. 1). These were possibly ethnolinguistic (less probably socio- 
economic) population groups. The total size of the Hephthalite population is unknown, 
but in Tokharistan alone there were 5-6,000 Hephthalite warriors - with their families, 
this suggests some 50,000 individuals, but there must have been considerable fluctuations 
during the period. 

Political and military history 

The political history of the Hephthalites can be deduced from Armenian, Arabic, Persian, 
Byzantine, Chinese and other sources. Under Yazdgird II (438-457) the north-eastern bor- 
ders of the Sasanian Empire were under threat from Central Asian tribes. The fifth-century 
Armenian historian, Elishe Vardapet, who was a contemporary, reports that the emperor 
was obliged to do battle with the tribes of the Hephthalites from 442. The situation was so 
serious that Yazdgird even had to transfer his residence to the northern border. It has been 
suggested that this was also the time when the Hephthalites made their appearance. 8 As 
early as 456, an embassy from the Hephthalites arrived in China. 9 

According to the Arab historian al-Tabari, Peroz, while still a prince, fled to the 'coun- 
try of the Haitals, or Hephthalites' and asked the king to provide him with troops to 'take 
possession of the kingdom of his father [Yazdgird II]'. Another source states that Peroz 
'was supported by the inhabitants of Tokharistan and the neighbouring regions' and refers 
to 'the people which conquered Tokharistan called Haital [that is Hephthalites]'. 10 In the 
mid-fifth century the Hephthalites increased greatly in strength and Tokharistan, with the 
surrounding regions, came under their rule. According to Harmatta, 'it is likely that the 
Hephthalites attacked the Transox[an]ian territory of the Kidarites in 466' and at the same 
time they 'took possession of the eastern part of Kusansahr, and then very soon they occu- 
pied also Balx [ Balkh] from the Persians'. 11 

7 Chavannes, 1903, pp. 229-33; Bailey, 1931, pp. 585-6; 1954, pp. 13-20. 

8 Trever, 1954, pp. 136-7; Gafurov, 1972, p. 198. 

9 Enoki, 1955, p. 234, Table. 

10 Shmidt, 1958, pp. 447-8; Noldeke, Tabari, 1973, pp. 117-19. 

11 Harmatta, 1969, p. 394. 



ISBN 978-92-3-103211-0 Contents Political and military history 

In gratitude for their assistance, Peroz extended the power of the Hephthalites still fur- 
ther. In particular, he ceded to them the district of Taliqan. 12 But there were disagreements 
between the Central Asian tribes and the Sasanians, leading to conflict and wars. First, 
Peroz clashed with the Hephthalites, who are considered by some scholars to have taken 
advantage of the civil war in Mesopotamia to seize Balkh (Priscus of Panion, 35). 13 There 
is, however, some doubt about this. What is known is that in the 460s or the 470s Peroz 
waged three wars against the Hephthalites. The first war ended in his being taken pris- 
oner and later released for a ransom partly paid by the Byzantine emperor. The second war 
ended as ingloriously as the first: Peroz was defeated and was once more taken prisoner. He 
was forced to give assurances never again to oppose the Hephthalites and to send instruc- 
tions that a huge ransom should be paid for him. Since the treasury was unable to send the 
ransom, Peroz left his son as hostage. 

The Hephthalites had strong forces. Sources describe them as skilful warriors and their 
army as powerful. They were armed with clubs and the Chinese considered that they were 
excellent archers. According to other sources, their main weapon was the sword. Judg- 
ing by their military operations, they probably possessed a strong cavalry force led by an 
asbarobido (cavalry commander). 14 

In Iran, according to Lazar of P'arp: 

Even in time of peace the mere sight or mention of a Hephthalite terrified everybody, and 
there was no question of going to war openly against one, for everybody remembered all too 
clearly the calamities and defeats inflicted by the Hephthalites on the king of the Aryans and 
on the Persians. 

Not only the common soldiers but also the dignitaries and military chiefs feared the 
Hephthalites. When Peroz set off on campaign, 'his troops went forward more like men 
condemned to death than warriors marching to war'. When news of the third campaign 
reached the king of the Hephthalites, he sent his representative to Peroz with this message: 
'You concluded peace with me in writing, under seal, and you promised not to make war 
against me. We defined common frontiers not to be crossed with hostile intent by either 
party.' 15 An important point to emerge from this text is that the Hephthalites appear not 
merely as a group of nomadic tribes but as a state formation, on an equal footing with 
Sasanian Iran and fully versed in statesmanship. 

Al-Tabari's text is very similar in this respect, although he incorrectly calls these tribes 
Turks (instead of Hephthalites). According to his account (which, however, also contains 

12 Balcami, 1869, pp. 127-8. 

13 Marshak, 1971, pp. 63-4. 

14 Altheim, 1960, Vol. 2, p. 269. 

15 Ter-Mkrtichyan, 1979, pp. 55-6. 



ISBN 978-92-3-103211-0 Contents Political and military history 

some elements of legend), Peroz reached the tower which had been built by Bahram Gur 
( Bahram V, 420-438) on the border between the regions of Khurasan and the Hephthalites 
to prevent them crossing into Khurasan; this was in accordance with the pact concluded 
between the Hephthalites and the Persians (i.e. Sasanians) that neither party should violate 
the border. Peroz, for his part, had promised Akhshunvar, the king of the Hephthalites, that 
he would not go beyond their borders. Peroz had 50 elephants and 300 men harnessed to 
Bahram Gur's tower. They drew the tower along in front of him while he marched behind, 
declaring that in that way he was not breaking his pact with Akhshunvar. 

The Hephthalite troops retreated, but then Akhshunvar ordered deep pits to be dug, 
lightly timbered over and topped with soil. These booby traps, laid in the path of the pursu- 
ing Sasanian army, played a decisive part, breaking its battle formation and ensnaring many 
soldiers. Peroz was killed and many of his retinue, including his daughter, were taken pris- 
oner by the Hephthalites, who seized his treasure. One of Peroz' followers, called Sukhra, 
subsequently managed to retaliate and forced the Hephthalites to withdraw. 16 A similar 
account is given by al-Dinawari. 17 

Following internecine conflict over the Sasanian throne, one of Peroz' sons, Kavad, 
fled to the Hephthalites. Having lived with great honour among them for four years, he 
married the daughter or sister of the Hephthalite king, who provided him with troops. 
Kavad seized the throne with these troops in 488, becoming king of Sasanian Iran. 18 As 
a result of political and kinship ties, the Hephthalites subsequently took part in Kavad Fs 
military campaigns and Hephthalite troops armed with cudgels were present at the siege of 
Edessa. 19 

As a result of internal events in Iran - the Mazdakite movement (see Chapter 17, 
Part One) and the revolt of the nobility against the king - Kavad once more fled to the 
Hephthalites. The Hephthalite king agreed to provide him with 30,000 troops; in return, 
Kavad was obliged to make territorial concessions and in 498 he handed over Chaganiyan 
to his allies. 20 Iran had to pay tribute to the Hephthalites for many decades, from 484 until 
the middle of the sixth century. Part of the Sasanian coinage was countermarked with a 
Hephthalite sign, and these were the coins used for payment of the tribute. 21 This situation 
continued into the early years of the reign of Khusrau I (531- 579). 

16 Shmidt, 1958, p. 449; Noldeke, Tabari, 1973, pp. 122-34. 

17 See Altheim, 1960, Vol. 2, pp. 51-2. 

18 Noldeke, Tabari, 1973, pp. 135-7. 

19 Pigulevskaya, 1941, p. 64. 

20 Shmidt, 1958, pp. 476-7. 

21 Gobi, 1967, pp. 193-4; 1971, p. 70. 



ISBN 978-92-3-10321 1-0 Contents Conquests in Gandhara and northern India 

In the second half of the fifth century and the first half of the sixth, silver coins of Peroz 
(and imitations with various overstrikes) circulated in northern Tokharistan. The genuine 
Peroz drachms belong mainly to the time when some of those regions, particularly Cha- 
ganiyan (Shi-han-na in the Chinese sources), were still under the Sasanians. Peroz drachms 
were subsequently minted in Chaganiyan with Bactrian and Sogdian overstrikes, together 
with imitations of Peroz coinage with Bactrian legends. Many coins were issued under 
Khusrau I, particularly from the 540s onward. Coins were subsequently minted which were 
imitations with overs truck names and portraits of local Hephthalite leaders. This series is 
completed with the issue of a coin of the type of Khusrau I, but with the name of a local 
ruler. 22 

The Hephthalites thus entered the historical arena in the mid-fifth century, apparently 
in eastern Tokharistan. By the end of the century they had taken possession of the whole 
of Tokharistan, including the Pamirs, and a considerable part of Afghanistan. At the same 
time, they seized much of East Turkestan. In 479 they subjugated the region of Turfan, 
and between 497 and 509 the region of Karashahr and what is today Urumchi. In 522 
P'o-lo-men, the leader of the Juan-juan in an area to the north of Dunhuang, fled to the 
Hephthalites to seek their protection. Earlier, probably in the late fifth century, Kashgar 
and Khotan had come under the power of the Hephthalites, who subjugated practically the 
whole of East Turkestan. As Enoki has correctly pointed out, the Hephthalites reached the 
zenith of their power with their seizure in 509 of Sughd (the capital of Sogdiana), which 
then ceased sending embassies to China. 

Conquests in Gandhara and northern India 

The late fifth and early sixth centuries saw the start of Hephthalite raids on Gandhara and 
subsequently on the whole of northern India. In 477 the Kidarites in Gandhara had sent 
an embassy to China, 23 but the Chinese pilgrim Sung Yiin, who visited Gandhara in 520, 
noted that the Hephthalites had conquered the country and set up their own ruler. 'Two 
generations then passed.' 24 On this basis, Marshall assumes that the invasion took place 
earlier (reckoning one generation to be 30 years: 520— 60=460). 25 

In the early second half of the fifth century the Hephthalites came into conflict with the 
Guptas, who had by then passed the zenith of their power and prosperity. According to the 
Junagadh rock inscription of c. 457, King Skandagupta won a victory over hostile kings 

22 Rtveladze, 1983, pp. 74-5. 

23 Enoki, 1955, pp. 234-5; 1959, pp. 25-7. 

24 Yang Hsiian-chih, 1984, p. 235. 

25 Marshall, 1951, p. 75. 



ISBN 978-92-3-10321 1-0 Contents Conquests in Gandhara and northern India 

and over tribes which seem to have been Hephthalites (or Kidarites; see pages 123-4). 
Another inscription proclaims that he bore the title, 'lord of a hundred kings'. 26 The initial 
attacks of the Hephthalites were thus repulsed, but at the price of stretching all the forces 
of the Gupta Empire. 27 

Skandagupta was the last great ruler of that dynasty, reigning from c. 454 to c. 467. 28 
The central power of the Gupta state subsequently declined, particularly in the last quarter 
of the fifth century under King Budhagupta, 29 the time at which the penetration of the 
Hephthalites into the subcontinent began. In the late fifth and early sixth centuries the 
Hephthalites in India came under the leadership of Toramana (see also Chapter 7, Part 
One), described in one Indian inscription as the 'renowned Toramana, the boundlessly 
famed ruler of the earth'. Launching an offensive from Panjab, he conquered the whole of 
western India and even Eran (in modern Madhya Pradesh). Numismatic evidence indicates 
that he ruled in Uttar Pradesh, Rajputana, Panjab and Kashmir. 30 His conquests brought 
with them the destruction of towns, villages and Buddhist monasteries, and the monasteries 
never recovered. Many local rulers acknowledged themselves to be subjects of Toramana. 31 

In the time of Toramana, the Hephthalites in India began to operate independently 
of the Central Asian branch, though the link between them does not seem to have been 
broken. 32 According to the Gwalior inscription, Toramana's son was called Mihirakula 
(in Jain sources, Caturmukha-Kalkin or Kalkiraja). He intensified his father's efforts to 
conquer the whole of northern India, and in this he was highly successful. 33 Over a cen- 
tury later, the Chinese pilgrim Hsiian-tsang paid particular attention to this ruler's life and 
activities in the account of his travels. He writes of Mihirakula: 'He was of quick talent and 
naturally brave. He subdued all the neighbouring provinces without exception.' 34 

The account of Cosmas Indicopleustes (writing in the early sixth century) confirms 
that the Hephthalites in India reached the zenith of their power under Mihirakula, with 
their capital at Sakala (modern Sialkot). Hsiian-tsang recounts the fate of Mihirakula: he 
was ultimately opposed by the Gupta ruler Baladitya, who had previously been paying him 
tribute (this is assumed to have been Narasimhagupta I). 35 Baladitya's opposition stemmed 

26 Fleet, 1888, pp. 14-15. 

27 Altekar and Majumdar, 1954, pp. 163-4; Majumdar, 1954, pp. 26-7. Concerning the Huns in Gupta 
inscriptions, see Sharma, 1978, pp. 133-4; Biswas, 1973. 

28 Gupta, 1974, pp. 329-38; cf. Singh and Bannerji, 1954, p. 89. 

29 Bongard-Levin and Il'in, 1969, p. 513. 

30 Majumdar, 1954, p. 35. 

31 Fleet, 1888, pp. 88, 159. 

32 Harmatta, 1969, pp. 400-1. 

33 Pathak, 1917, pp. 215-16; Gupta, 1974, pp. 372-7. 

34 Beal, 1969, p. 167. 

35 See Gupta, 1974, pp. 364-6. 



ISBN 978-92-3-10321 1-0 Contents The Hephthalites of Central Asia 

from the atrocities perpetrated by the Hephthalite leader and the destruction of Buddhist 
buildings, which is also reported in the Kashmir chronicle the Rajataranginn and in Jain 

According to Hsiian-tsang, Mihirakula was taken prisoner by Baladitya, but was subse- 
quently released. Power over the Hephthalite tribes had meanwhile been seized by Mihi- 
rakula's brother and Mihirakula himself set off for Kashmir, where the king received him 
with honour. A few years later, Mihirakula incited the townspeople of Kashmir to revolt 
against their king and seized power there. He then went westwards and occupied Gand- 
hara, where he killed many of the inhabitants and destroyed the Buddhist shrines, only to 
die shortly afterwards. 36 While the details of this account by Hsiian-tsang may be unhis- 
torical, the broad outline is worthy of note. 

In assembling the events of Mihirakula's life, the RajataranginT asserts that he was a 
powerful king who ruled Kashmir and Gandhara and even (this is clearly an exaggeration) 
conquered southern India and Ceylon. Cosmas Indicopleustes calls him 'king of India', 
though he mentions that the possessions of the Huns in India (i.e. Hunas) were divided 
from the other Indian kingdoms by the mighty River Phison (Indus). 

Persecution of the local population, combined with religious intolerance, set the local 
Indian population against the Hephthalites and deprived them of support. At the same time, 
the difficulties facing the Hephthalites of Central Asia in their struggle against the Turks, 
who utterly defeated them, deprived the Hephthalites in India of their Central Asian base, 
of the 'flow of fresh forces and support, and this led to their decline'. 37 Although small 
communities and even principalities of Hephthalites survived in India after the middle of 
the sixth century, they did not wield any significant political influence. 

The Hephthalites of Central Asia 

By the middle of the sixth century, the Hephthalites of Central Asia found themselves 
squeezed between Sasanian Iran, whose power had increased tremendously under Khusrau 
I, and the Tiirks, who had conquered much of the north-east of Central Asia. The oppo- 
nents of the Hephthalites entered into diplomatic negotiations with one another, but when 
the kaghan of the Tiirks dispatched ambassadors to Iran, they were killed in Hephthalite 
territory at the command of the Hephthalite king. The kaghan moved his forces and seized 
Chach (modern Tashkent) and continued to the Syr Darya (Jaxartes). The forces of the 
Hephthalites gathered in the region of Bukhara, towards which Hephthalite detachments 

36 Beat, 1969, pp. 169-72. 

37 Gafurov, 1972, p. 201; see also Majumdar, 1954, p. 39. 



ISBN 978-92-3-103211-0 Contents Social structure and administration 

marched from Termez, southern Tajikistan and even the Pamirs. An eight-day battle was 
fought in the Bukhara area, in the course of which the Hephthalites were routed. Their 
troops fled south and there elected a new king, Faganish (or Afganish), but the south of 
Central Asia had been occupied by Sasanian troops and the new Hephthalite ruler acknowl- 
edged the supremacy of Khusrau I (see also Chapter 7). This marked the end of the Heph- 
thalite state in Central Asia. (These events took place in the period 560- 563. ) 38 

Central Asia was devastated as a result of this struggle, whereupon relations between 
the allies ( Turks and Sasanians) became strained. This worked to the advantage of the 
Hephthalites: individual semi-independent Hephthalite principalities continued to exist in 
the Zerafshan valley, paying tribute to the Turks ( Menander, fragment 18). The situation 
was similar in the south, except that here the Hephthalites paid tribute to the Sasanians. 
Khusrau I found a pretext to cross the Amu Darya (Oxus). Power over the littoral of the 
Amu Darya later passed to the Turks, who then occupied all the territory of Afghanistan. 
Small Hephthalite principalities continued to exist in southern Tajikistan and Afghanistan 
for a long time; some of them (in particular Kabul) remained independent. 39 

According to Gafurov: 

The Hephthalites thus established a huge state structure even greater in geographic extent 
than that of the Kushans, but at the same time it was more loosely-knit and more unstable. 
They succeeded both in halting the armies of Sasanian Iran in the east and in inflicting a 
shattering defeat on the Sasanian kings. Hephthalite rulers even settled succession claims to 
the title of shahanshah of Iran, while regular payment of tribute to them was a major concern 
for many Iranian governments. In conclusion, the Hephthalites played an important part in 
the ethnogenesis of the peoples of India, Afghanistan and, in particular, Central Asia. 40 

Social structure and administration 

Although some evidence remains of the society of the Hephthalites, their customs and ways 
of living, the information is extremely contradictory. According to Procopius, the Heph- 
thalites had 'since time immemorial' lived a settled life, 'were ruled by one king' and 'had 
a state system based on law'. A Turk mission reported to Byzantium that the Hephthalites 
were 'a tribe which dwelt in towns' (see also pages 149-50); indeed, after their victory 
over the Hephthalites, the Turks became 'the masters of their towns' (Menander, fragment 
18). Theophanes Homologates (fragment 3) states that after their victory over the Irani- 
ans, the Hephthalites became the masters of the towns and ports previously held by their 

38 Mohl, 1868, pp. 306-16; Shmidt, 1958, p. 453; Noldeke, Tabari, 1973, pp. 158-9; Moravcsik, 1958, pp. 
275-6; Grignaschi, 1980. 

39 Harmatta, 1969, p. 402. 

40 Gafurov, 1972, p. 202. 



ISBN 978-92-3-103211-0 Contents Social structure and administration 

enemy. 41 Chinese chronicles and travellers, however, provide a different picture. For exam- 
ple, according to the Sui shu [Dynastic Annals of the Sui], the Pei-shih and other sources, 
in the land of the Hephthalites there were 'neither cities and towns nor fixed residence of 
their king'. 42 Sung Yiin reported in 518 that, in the land of the Hephthalites: 

there were no walled cities for residences; [the area] was kept in good order by a patrolling 
army. The people lived in felt [tents], moving from one place to another in pursuit of water 
and pasture lands: they moved to cooler areas in summer and warm regions in winter. The 
natives were simple rustic folk, unversed in writing the rites or moral precepts. 43 

Even the Chinese sources do not agree, however. Thus, according to the Chou shu [His- 
tory of the Chou Dynasty] (15a), 'it [the land of the Hephthalites] has its capital in the 
walled city of Pa-ti-yen', a name meaning something like 'the walled city in which the 
king resides'. 44 Hsiian-tsang's report on the country of Hsi-mo-ta-lo (a Sanskritized form 
of the ethnonym, Heptal) helps to resolve these conflicting accounts. According to the 
Chinese pilgrim, after their ancestors had established a strong state and subjugated their 
neighbours, the Hephthalites 'migrated and scattered in foreign countries where they rule 
scores of strongly walled cities and towns with so many chiefs. They [also] live in tents of 
felt and remove from one place to another.' 45 

The following explanation may be advanced for all the differing accounts. The core of 
the Hephthalite population was originally nomadic or semi-nomadic but later, after seizing 
control of vast regions with towns and fortresses, the Hephthalite elite, like other conquer- 
ing nomads (for example, the Karakhanids and the Seljuks), began to settle in towns and 
were followed by other groups from the newly arrived non-urban population. The sources 
simply mention isolated episodes in this complex story - hence the discrepancies. 

One of the most difficult questions concerns the Hephthalites' social structure. From the 
description of their funeral rites in the Chinese chronicles (see pages 147-8), we learn that 
there were both rich and poor Hephthalites and that their rites were completely different. 46 
Consequently, it was a class society with marked social and property differentiation. In 
his description of the luxurious dwelling of the Hephthalite king with his golden throne, 
and the magnificent clothes of the king and queen inlaid with gold and precious stones, 
Sung Yiin notes that 'there were differences, it was observed, between the nobleman and 

41 For a detailed analysis of Byzantine sources, see Moravcsik, 1958. 

42 Enoki, 1959, p. 10. 

43 The traveller, of course, meant Chinese writing, rites and moral precepts. See Yang Hsiianchih, 1984, p. 

44 Miller, 1959, p. 12. 

45 Enoki, 1959, p. 35. 

46 Ibid., p. 49. 



ISBN 978-92-3-103211-0 Contents Social structure and administration 

commoners'. He continues: 'For the people's clothing and ornaments, there was nothing 
but felt.' 47 

At the apex of Hephthalite society was the king, whose residence was a fortified town. 48 
According to Byzantine sources, the Hephthalites 'were ruled by one king' . The legends on 
coins sometimes contain the terms XAHO and XOAAHO (sovereign) and the expression 
'great sovereign' is occasionally encountered. 49 Names of individual monarchs are known 
(some from historical accounts). Thus, according to Firdausi, the Hephthalites were led by 
a king called Gatfar during their struggle against the Turks, which ended in their defeat 
at the battle near Bukhara (see page 143 above). It is possible that the kings were chosen 
in peacetime as well as in exceptional circumstances but it is not known who chose them, 
perhaps the elite. One Chinese account states that the throne of the Hephthalites 'was not 
transmitted by inheritance but awarded to the most capable kinsman'. 50 

The Hephthalite state covered a huge territory and the regions forming it were dependent 
upon the central authority to varying degrees. According to Sung Yun, 'the state received 
tribute from a number of countries. . . altogether delegates from more than forty countries 
came to pay tribute and offer congratulations on appropriate occasions'. 51 According to 
another source, countries 'large and small, altogether more than twenty, are all subject to 
it [the Hephthalite state]'. 52 

There was an administrative machinery at both central and regional levels. During the 
550s the Hephthalite king had an adviser (minister?) by the name of Katulf ( Menander, 
fragment 10). Such titles as oazorko,fromalaro, hazaroxto and asbarobido (commander of 
the cavalry) are known from inscriptions on gemstones. 53 The state system was a complex 
amalgam of institutions originating in Hephthalite society and frequently going back to 
ancestral tribal arrangements, as well as institutions which were native to the conquered 
regions. Money was minted and we know of many series of coins. Excellent classificatory 
and typological works have been published, 54 but insufficient use has been made of these 
coins as a historical source. 

Central control in the Hephthalite state was weak and local dynasties continued to 
rule in a number of regions. Such was the case in Chaganiyan, on the upper and middle 
reaches of the Surkhan Darya. One of the rulers of this dynasty was Faganish (see page 143 

47 Yang Hsiian-chih, 1984, pp. 224-6. 

48 Miller, 1959, p. 12. 

49 Livshits, 1969, pp. 69-71. 

50 Bichurin, 1950, p. 269. 

51 Yang Hsiian-chih, 1984, p. 225. 

52 Miller, 1959, p. 21. 

53 Livshits, 1969, pp. 66-7. 

54 Gobi, 1967. 



ISBN 978-92-3-103211-0 Contents Religious life and polyandry 

above), whose name is known from written sources; the names of other rulers appear on 
Chaganiyan-Hephthalite coins. The name of another Chaganiyan ruler, Turantash, appears 
on a long inscription at Afrasiab. In the first quarter of the eighth century, Chaganiyan was 
ruled by Tish, the 'One-Eyed' (in the Sogdian language, Tish is the name of the star Sirius). 
The Manichaean religion was widespread in Chaganiyan together with Buddhism. 55 

Another powerful domain, Khuttal (Kou-tou or Kou-tou-lo in the Chinese sources), was 
also associated with the Hephthalites. It was located in the basin of the River Kyzyl-su, but 
at times also included the basin of the River Vakhsh. Khuttal also had a local dynasty with 
an established order of succession to the throne, according to Arabic sources. The rulers 
took the Iranian title of Khuttal- shah or sher-i Khuttal. The Arabs referred to them as muluk 
(pi. of malik, king). 56 In the southern part of Central Asia and northern Afghanistan, in the 
region known as Bactria under the Achaemenids and later as Tokharistan (T'ou-ho-lo or 
Tou-ho-lo in Chinese sources from 383), there were some 30 dominions in the sixth to the 
seventh century with their own rulers, some of whom were of Hephthalite extraction. 

Religious life and polyandry 

Information about the religion of the Hephthalites is provided by the Chinese sources. Sung 
Yiin reports that in Tokharistan 'the majority of them do not believe in Buddhism. Most of 
them worship wai-shen or "foreign gods".' He makes almost identical remarks about the 
Hephthalites of Gandhara, saying that they honour kui-shen (demons). The manuscripts 
of the Liang shu (Book 54) contain important evidence: '[the Hephthalites] worship T'ien- 
shen or [the] heaven god and Huo-shen or [the] fire god. Every morning they first go outside 
[of their tents] and pray to [the] gods and then take breakfast.' For the Chinese observer, 
the heaven god and the fire god were evidently foreign gods. We have no evidence of the 
specific content of these religious beliefs but it is quite possible that they belonged to the 
Iranian (or Indo-Iranian) group. 57 

Although Sung Yiin states that the Hephthalites did not believe in Buddhism, Bud- 
dhist religious establishments flourished in Tokharistan and other areas. In India, however, 
the Hephthalites showed intolerance towards Buddhist religious establishments. It may be 
supposed that the beliefs of the local subject populations - including Buddhism - grad- 
ually began to gain ground among the Hephthalites. Various forms of Zoroastrian beliefs 
were widespread in Central Asia and northern and western Afghanistan in competition 

55 Rtveladze, 1973; Bosworth, 1981. 

56 Marquart, 1901, p. 30; Belenitskiy, 1950, p. 115. 

57 Enoki, 1959, pp. 45-9. 



ISBN 978-92-3-103211-0 Contents Language and scripts 

with Buddhism. There were also many adherents of Hindu beliefs in Afghanistan and in 
Tokharistan. Lastly, Manichaeism had taken firm root and Christianity was spreading. 

Chinese sources provide the following account of the funeral rites mentioned above: 
'if a man dies, a wealthy family will pile up stones to form a house [to keep the corpse]; 
a poor family will dig the ground for burial. The articles of everyday use are buried with 
the dead.' Another source describes a third type of burial: 'In burying the dead, the cof- 
fin is laid in a wooden case. When a parent dies, the child will cut off one of his ears.' 58 
It is known, however, that various types of burial structure, including small, surface-level 
stone houses, pit graves and wooden coffins, were employed at the same period in Fer- 
ghana; hence the hypothesis that these Chinese accounts are actually descriptions of life in 
Ferghana. 59 

Polyandry was the Hephthalites' most noteworthy social custom. Brothers had one wife 
in common and the children were considered as belonging to the oldest brother. The num- 
ber of 'horns' on a married woman's headdress corresponded to the number of her hus- 
bands. This custom was practised in ancient times among the Central Asian Saka people, 
the Massagetae (Herodotus, I, 216); in medieval Afghanistan (according to al-Biruni); and 
among present-day Tibetans. 60 

Language and scripts 

Two accounts of the Hephthalite language have been quoted above, but they are not very 
informative. Some scholars believe that the Hephthalites spoke a Turkic language while 
others affirm that their language belonged to the East Iranian group. Although the reading 
of Hephthalite coins and gemstones is still the subject of much controversy, an 
interpretation of the names and titles appearing on them, and of the names of Hephthalite 
rulers in Firdausi's Shah-name, is possible from East Iranian languages. This does not, 
however, constitute a conclusive argument. At present it can only be asserted that Bactrian 
enjoyed the status of an official language in the Hephthalite domains in Tokharistan. 61 

Late Bactrian script is a development of Bactrian script, which was itself an adapted 
form of the Greek alphabet. Hephthalite script is typically semicursive or cursive and much 
of it is difficult (or impossible) to read. Examples of the Hephthalite written language 
have been discovered in East Turkestan, Central Asia, Afghanistan and northern Pakistan. 
However, these are only insignificant vestiges of the large quantity of written material 

58 Enoki, 1959, pp. 49-50. 

59 Litvinsky, 1976, p. 56. 

60 For details, see Enoki, 1959, pp. 51-6. 

61 Livshits, 1969, p. 67. 



ISBN 978-92-3-103211-0 Contents Towns 

which, if we are to believe Hsiian-tsang, was to be found in the regions occupied by the 
Hephthalites and particularly in Tokharistan. 62 

In the vast region controlled by the Hephthalites, people spoke various languages, 
including Iranian (Middle Iranian, especially Bactrian, but also Sogdian and Middle Per- 
sian, or Pahlavi), Indian tongues and others of which we have no written records. Various 
scripts were also in use, particularly Bactrian, Pahlavi, KharosthT and Brahmi. 


The Hephthalite economy was composed of three sectors: urban, settled agricultural and 
nomadic. Urban settlements did not outnumber rural settlements, yet the economic, polit- 
ical, religious and cultural role of the towns was far more important than that of the vil- 
lages. Although very little is known about the towns during the fifth and sixth centuries, it 
has been established that one of the largest towns was Balkh, where exploratory excava- 
tions have been undertaken. Hsiian-tsang (writing in 629) describes Po-ho (Balkh) as the 
Hephthalite capital, with a circumference of approximately 20 li. He continues: 'This city, 
though well [strongly] fortified, is thinly populated.' Balkh had about 100 Buddhist vihdras 
(monasteries) and some 3,000 monks. Outside the town was a large Buddhist monastery, 
later known as Naubahar. 63 Some idea of what Balkh looked like in the fifth and sixth 
centuries may be obtained from the descriptions of Arab authors, 64 but their accounts all 
date from a later period. Unfortunately, little archaeological work has been carried out 
on Balkh. 65 

Of the same size as Balkh was the early medieval town of Termez which, according to 
Hsiian-tsang, lay on an east-west axis and had a circumference of about 20 li. Termez had 
some 10 sarigharamas (monasteries) and perhaps 1,000 monks. 66 Excavations have been 
conducted there but little evidence has been found of the town between the fifth and the 
seventh century. It consisted of a rectangular shahristan, or town (roughly 10 ha in area), 
and a large suburb enclosed by a wall. The total area was approximately 70 ha and the 
entire town was probably surrounded by a wall about 6 km long. It is likely that there was 
also a citadel. 67 

62 Steblin-Kamenskiy, 1981. 

63 Beal, 1969, pp. 43-6; cf. Hui-li, 1959, pp. 50-2. 

64 Schwarz, 1933, pp. 434-43. 

65 See Le Berre and Schlumberger, 1964; Young, 1955. 

66 Beal, 1969, Vol. 1, pp. 38-9. 

67 Shishkin, 1940, pp. 150-1; Belenitskiy et al., 1973, pp. 177- 



ISBN 978-92-3-103211-0 Contents Architecture 

According to Hsiian-tsang, the capital of Chaganiyan was half the size of Termez and 
Balkh in terms of its circumference (10 li) and had five Buddhist monasteries. 68 It has been 
identified with the site of Budrach, which even in Kushan times had an area of 20 ha and 
at the period under consideration occupied a much greater area than the Kushan town. The 
expanded town had a rectangular citadel, a fortified shahristan with an area of over 50 ha 
and, beyond that, a large suburban area with farms, forts and religious edifices. 69 

The capitals of other regional domains were roughly similar to or larger than the capital 
of Chaganiyan. The capital of the province of Hu-sha (or Vakhsh) had a circumference 
of 16-17 li 10 and is the site of Kafyr-kala in the Vakhsh valley. It has a walled citadel 
(measuring 360x360(m) in one corner of the rectangular town, which is, like the citadel, 
surrounded by a wall with towers. The citadel contained the palace of the ruler (see below). 
The town was divided in half by a central thoroughfare on which stood dwellings, and reli- 
gious and commercial buildings. Outside the town fortifications lay extensive suburbs. 71 
There were also medium- and small-sized towns such as Kala-i Kafirnigan. 72 


An idea of the architecture of Tokharistan at the time of the Hephthalites is provided by 
the palace (KF-II period) in the citadel of Kafyr-kala (see Chapter 7). It had a square 
plan (70x70 m) and was encircled by two walls, a main inner wall and a secondary outer 
one (proteikhos). Strong rectangular towers were located at the corners, and in the cen- 
tre of the wall stood semi-circular projecting towers. Between the corner towers and the 
projecting towers were recessed bays with arches containing false arrow slits. The passage- 
ways running along the protective walls had a defensive function. 

In the courtyard of the citadel stood the palace buildings. The palace was laid out around 
a central rectangular hall with an area of 200 sq. m and was surrounded by smaller halls and 
domestic buildings. In the north-west corner stood a circular hall (8 m in diameter) with a 
corbelled cupola. This hall was entered from the east side. In the south-east corner was a 
Buddhist vihdra, a square sanctuary whose walls were decorated with polychrome murals 
depicting the Buddha and other Buddhist motifs. Buildings serving various purposes were 
located on the square of the shahristan. One complex has been excavated consisting of a 
central rectangular hall with a bay flanked by columns in the rear wall, and by subsidiary 

68 Beat, 1969, Vol. 1, p. 39. 

69 Rtveladze, 1983. 

70 Beal, 1969, Vol. l,p. 40. 

71 Litvinsky and Solov'ev, 1985. 

72 Litvinsky, 1981. 



ISBN 978-92-3-103211-0 Contents Art and crafts 

rooms. 73 Whereas the palace had only one storey, there were also two-storey buildings such 
as Kuev-kurgan near Termez which, on the second floor, had a frieze of painted statues. 

Some of the Buddhist buildings were extremely large, such as the 'New Sarigharama' 
near Balkh, which contained a magnificent statue of the Buddha in a spacious room. Both 
the statue and the room were decorated with rare and precious substances. There was also 
a renowned statue of the deity Vaishravana Deva. 'To the north of the convent is a stupa, 
in height about 200 feet [61 m], which is covered with a plaster hard as the diamond, and 
ornamented with a variety of precious substances.' 74 The tenth-century Arab geographer 
Ibn al-Faqih calls the main building al-asbat: the diameter and height of the stupa were said 
to be 100 cubits (about 35 m). It was surrounded by porticoes and contained 360 separate 

The principal building materials were large, rectangular, unbaked bricks and large pakhsa 
blocks (made of clay mixed with finely chopped straw). The walls were often composite, 
consisting of pakhsa blocks with intervening layers of brick, or else the lower part was 
made of pakhsa and the upper part of brick. Little use was made of baked brick but wood 
(for columns, and so on) was often employed. The roofs were either flat (often supported 
by columns) or arched and domed. The arches were faced with sloping segments while the 
domes were either corbelled (and formed of a horizontal brick overhang) or supported by 
squinches. 75 In areas south of the Hindu Kush mountains such as Kapisa and Gandhara, 
the main building material was stone. 

Art and crafts 

Several murals at Dilberjin near Balkh date from the fifth to the seventh century. In the 
centre of one of these murals is a large female figure sitting on a throne with her knees 
wide apart. In her left hand she holds a shield, and on her head is an intricate headdress 
reminiscent of a helmet. Hornlike projections rise above her shoulders and behind her head 
is a halo. She has ornaments on her neck and hands; a cloak is thrown over one shoulder 
and she is wearing a belt round her clothing. Two smaller figures, one very small, are 
moving from the sides towards the central figure. Other murals feature a procession of 
men standing en face, some armed with daggers. They are dressed in narrow kaftans with 
turned-down flaps on the right-hand side. A distinctive feature of these figures is their 
position 'on tiptoes'. Another scene shows a feast with sitting or semi-reclining figures 

73 Litvinsky and Solov'ev, 1985, pp. 8-95. 

74 Beat, 1969, Vol. 1, pp. 44-8. 

75 Litvinsky and Solov'ev, 1985, pp. 49-85; Pugachenkova, 1976, pp. 157-62. 



ISBN 978-92-3-103211-0 


Art and crafts 

holding goblets in their hands. There are also elaborate compositions whose significance is 
not yet clear. 76 A comparison between some of the Dilberjin paintings and those at Kyzyl 
(see illustrations in Chapter 11), particularly 'the cave of the 16 swordsmen' and 'the cave 
with the picture of Maya' , demonstrates the link between them and enables us to date these 
Dilberjin paintings to the fifth or early sixth century. 77 

The remarkable cycle of paintings at Balalyk-tepe depicting a feast dates to the end of 
the sixth or beginning of the seventh century. Great lords are seated with beautiful ladies, 
drinking wine from golden goblets while servants hold large umbrellas over them. The 
garments and ornaments of the figures at Balalyk-tepe are similar to those of the figures at 
Dilberjin. 78 

FIG. 2. Bamiyan. Panoramic view of the complex. (Photo: © Andrea Bruno.) 

76 Kruglikova, 1976; 1979; see also Buriy, 1979 

77 Litvinsky and Solov'ev, 1985, p. 139. 

78 at 

Al'baum, 1960; on the dating, see Antonini, 1972, pp. 71-7; Belenitskiy and Marshak, 1979, p. 35. 



ISBN 978-92-3-103211-0 


Art and crafts 

FIG. 3. Bamiyan. View of the stone cliff. (Photo: © Andrea Bruno.) 

An enormous quantity of artistic and architectural remains have been preserved in cen- 
tral and south-eastern Afghanistan, amongst which are the objects found at the world- 
famous complex in Bamiyan (Figs. 2 and 3). There, in a mountain valley, lived a 
population whose customs, literature and coinage, according to Beal, 'are the same as 
those of the Tokharistan country. Their language is a little different, but in point of per- 
sonal appearance they closely resemble each other.' 79 

It was at Bamiyan that Hsiian-tsang saw gigantic standing figures of the Buddha and 
10 Buddhist establishments with 1,000 monks. They belonged to the Hinayana and the 
Lokottaravadin schools. The Chinese traveller was also struck by the size of the figure of 
the Buddha reclining in Nirvana. Two huge standing figures of the Buddha carved in the 
stone cliff have been preserved, one 38 m in height (Fig. 4), the other 53 m (Figs. 5-7). 
At a distance of some 1,800 m, the cliff is pierced at different levels by Buddhist cave 
edifices (of which some 750 remain) in which splendid paintings (Figs. 8-11), mainly with 


Beat, 1969, Vol. l,p. 50. 



ISBN 978-92-3-103211-0 


Art and crafts 


i ••■Mil '-it,^ 

FIG. 4. Bamiyan. Figure of the 38-m-high Buddha and cave edifices. 

FIG. 5. Bamiyan. Figure of the 53-m-high Buddha. (Photos: © Andrea Bruno.) 

1 57 


ISBN 978-92-3-103211-0 


Art and crafts 

FIG. 6. Bamiyan. Close-up view of the highest Buddha. (Photo: © UNESCO/A. Lezine.) 

Buddhist motifs, remarkable high reliefs, and so on (Figs. 12-14), have been preserved. 
The entire complex dates from between the third and the seventh century. The large figure 
of the Buddha is probably linked to the Hephthalite period. 

Also dating from that period is an entire series of works of art, particularly from Balalyk- 
tepe and Kyzyl, in which Indian, Sasanian and Central Asian influences can be traced. Also 
worthy of mention are the complexes at nearby Kakrak (Figs. 15 and 16) and, much further 



ISBN 978-92-3-103211-0 


Art and crafts 

FIG. 7. Bamiyan. Close-up view of the drapery on the highest Buddha. (Photo: © Andrea Bruno.) 

FIG. 8. Bamiyan. Detail of the painted decoration in the cave of the highest Buddha. (Photo: © 
Andrea Bruno.) 

1 59 


ISBN 978-92-3-103211-0 


Art and crafts 


FIG. 9. Bamiyan. Detail of the painted decoration in the cave of the highest Buddha. (Photo: © 
Andrea Bruno.) 

FIG. 10. Detail of a wall-painting. Musee Guimet, Paris. Photo: © R.M.N./© Ravaux. 



ISBN 978-92-3-103211-0 


Art and crafts 

FIG. 11. Bamiyan. Wall-painting. Woman musician. Photo: © R.M.N./© Droits reserves. 

FIG. 12. Bamiyan. Detail of a relief laternendeke from cave 5. Musee Guimet, Paris. Photo: © 
R.M.N./© Droits reserves. 



ISBN 978-92-3-103211-0 


Art and crafts 

FIG. 13. Bamiyan. Interior laternendeke decoration from a cave. (Photo: © Andrea Bruno.) 

FIG. 14. Bamiyan. Interior laternendeke decoration from a cave. (Photo: © Andrea Bruno.) 



ISBN 978-92-3-103211-0 


Art and crafts 

FIG. 15. Kakrak. Wall-painting from the ceiling of a cave. Musee Guimet, Paris. Photo: © R.M.N./© 
Richard Lambert. 

FIG. 16. Kakrak. Statue of a Buddha. (Photo: © Andrea Bruno.) 

off, at Dukhtar-i Nushirvan, and their paintings. 80 Huge sculptures of the seated Buddha 
and Buddha in Nirvana have been excavated at the vast Buddhist monastery of Tepe Sar- 
dar in Ghazni, where the central stupa and many surrounding votive stupas and places of 
worship have been unearthed. 81 

80 Godard and Hackin, 1928; Hackin and Carl, 1933; Rowland, 1946; Tarzi, 1977; Higuchi, 1983-4, Vols. 
1-3; Kuwayama, 1987; Klimburg, 1987. 


Taddei, 1968; 1974. 



ISBN 978-92-3-103211-0 


Art and crafts 

FIG. 17. Chilek. Decorated silver bowl. (Photo: © Vladimir Terebenin.) 

FIG. 18. Chilek. Decorated silver bowl. (Photo: © Vladimir Terebenin.) 



ISBN 978-92-3-103211-0 Contents Art and crafts 

Alongside monumental art, 'chamber arts' such as toreutics and the modelling of fig- 
urines were highly developed during the Hephthalite period. In India and Tokharistan, tore- 
utic artists produced fine work, including silver (sometimes silver gilt) bowls depicting 
nude and semi-nude women: for example, dancers with scarves over their heads hanging 
down to their thighs at a royal feast, where the central element in the composition is the 
bust-length image of a king that resembles the portraits of kings on Hephthalite coins (see 
the Chilek bowl: Figs. 17 and 18). Episodes from the hunt are depicted on a bowl from 
Swat. The entire scene is executed with panache: horses race in a frenzied gallop; hunters 
and their prey are not only shown in movement but with complex foreshortening; and some 
episodes express the authentic drama of a real hunt. 82 

The modelling of figurines also achieved a high level. At the beginning of the Middle 
Ages, this art form was represented by applique work in the shape of human busts under 
the handles of vessels and, in later examples, under the spouts (Ak-tepe II, Tutkaul, and so 
on). It sometimes took the form of entire scenes pressed out of a mould and applied to the 
walls of large vessels or, at other times, separate ceramic tiles with impressed images. 83 

82 Dalton, 1964, pp. 53-5, 58-9, PI. XXIX-XXXIII; Marshak, 1986, pp. 29-31, Figs. 11-13; Yakubov, 
1985; Pugachenkova, 1987. 

83 Litvinsky and Solov'ev, 1985; Sedov, 1987, pp. 103-5. 



ISBN 978-92-3-103211-0 Contents EASTERN KUSHANS 



A. H. Dani, B. A. Litvinsky and M. H. Zamir Safi 



Eastern Kushans 167 

Kidarites in Gandhara and Kashmir 171 

Later Hephthalites 173 

Economic and cultural progress 176 

Sanskrit references 178 

Coinage of the Hunas 179 


The Khalaj, the successors of the Hephthalites 184 

Urban life and art in Tokharistan 186 

See Maps 2 and 4. 



ISBN 978-92-3-103211-0 Contents Eastern Kushans 

Part One 


(A H. Dani) 

Eastern Kushans 

The extensive empire of the Great Kushans lasted until the close of the reign of Vasudeva 
I in the early part of the third century, 1 but even during his rule there is little evidence of 
his continued control over the eastern parts of the Gangetic valley in northern India. 2 The 
campaigns of the Sasanian rulers Ardashir I and Shapur I (see Chapter 2) put an end to the 
prosperity of the Great Kushans. 3 On the evidence of the Begram excavations, this collapse 
was dated to 244; but the eminent Austrian and Russian numismatists, Gobi and Zeimal, 
put it much later. Neither date is generally accepted by Indian scholars, 4 who place the end 
of the Great Kushans Empire at c. 180. This date accords well with the contemporary rise 
of several tribal states in northern India. 5 These states ousted the Kushans and gradually 
usurped power right up to eastern Panjab, although in Gandhara and central and western 
Panjab the Later Kushans maintained their hold. Though weakened and hemmed in by the 
rise of new powers, the Kushans continued to exercise authority from Bactria to Panjab. 

The names of only two Kushan kings of this period are known from coins. Many his- 
torians take them to be Kanishka III and Vasudeva II and distinguish them on numismatic 
grounds from earlier Kushan rulers bearing the same names. Numerous coins of Kan- 
ishka III have been found in Panjab, Seistan (modern Sistan) and Afghanistan, and also in 

1 The exact date depends upon fixing the beginning of the Kanishka era (for discussion, see Volume II). 

2 Sastri, 1957, pp. 247-8. 

3 Ghirshman, 1954, p. 291; but see Chattopadhyay, 1979, p. 92, where he maintains that the successors of 
Vasudeva I ruled in Bactria and Afghanistan. 

4 Altekar and Majumdar, 1946, p. 12. 

5 Ibid.,Ch. 2. 



ISBN 978-92-3-103211-0 Contents Eastern Kushans 

southern Uzbekistan and southern Tajikistan. His coins bear other abbreviated names in 
Brahmi which are taken to be those of his satraps, or governors. 6 

Altekar cites evidence from the Puranas (popular Hindu texts in epic form, containing 
some historical material; see Chapter 8), 7 which speak of 8 Yavana, 14 Tushara and 13 
Murunda chiefs who probably ruled in Panjab in about the third and fourth centuries. These 
chiefs have been identified with names obtained from a series of coins found in Gandhara 
and Panjab. As all these coins follow the Ardokhsho type of Kanishka III, the names on 
them are thought to be the successors representing the Eastern Kushans, although not all 
of them may be ethnically related to the Kushans. In the excavation at Andan Dheri 8 four 
hoards of copper coins were found, all belonging to this period. On the obverse, they bear 
the name Saka in BrahmT. Together with the tribal name, shortened names of the chiefs 
occur, such as Sayatha, Sita, Sena, Pra, Mi, Shri and Bha. These are understood to be the 
names of seven Saka rulers who had authority over Gandhara 9 after Vasudeva II or III, and 
probably ruled for nearly 100 years. 

The Russian numismatist Zeimal, however, disagrees with the theory that there were 
several kings with the name of Vasudeva. He believes that there was only one king with 
this name who issued gold and bronze coins. Gold and bronze coins with the name of a 
later 'Kanishka' look like those of Vasudeva (though they are not identical). According to 
Zeimal, the comparison of gold coins of Kanishka in the Later Kushan series with those 
of Vasudeva, the last Great Kushan king, allows us to determine their relative chronology. 
The typological similarity of the earliest coins of the later Kanishka representing the king 
in armour allows us to classify them as gold coins of Vasudeva which formed the second 
stage in his minting. The similarity is so obvious that it allows us to regard them as issued 
at the same mint and probably by the same master, who then stopped making dies for 
the coins of Vasudeva. Thus the coins of the later Kanishka 'detach themselves' from the 
typological series of the gold coins of Vasudeva, and begin their own series. Rosenfield 
notes similar features in the gold coins of Vasudeva and those of the later Kanishka. 10 

Zeimal supposes that they reigned at the same period but in different territories, i.e. 
somewhere in the first half of Vasudeva's reign the kingdom of the Great Kushans was 
divided into two parts. A number of suppositions was made concerning the manner of the 
separation into western and eastern parts or north and south. As a rule, the starting point 
of the above discussions is a further development of the types of the enthroned Ardokhsho 

6 Banerji, 1908, p. 86; Altekar and Majumdar, 1946, p. 14. 

7 Pargiter, 1913, p. 45. 

8 Dani, 1968-69, pp. 42-6. 

9 Altekar and Majumdar, 1946, pp. 28-9; Chattopadhyay, 1979, pp. 102-3. 
10 Zeimal, 1983, p. 223. 



ISBN 978-92-3-103211-0 Contents Eastern Kushans 

(= the later Kanishka) and Shiva in front of the bull (= Vasudeva). The Ardokhsho type 
was further developed in the Gupta and Kidarite coinage, while the Shiva type was used as 
a prototype for the Kushano-Sasanian coins. The circulation areas of coins of these kings 
were different, but their distribution is poorly investigated. That is why Zeimal considers 
that the boundaries of the territories they each controlled should be left open so far. 11 

The Kushan coin type of the standing king on the obverse and the seated Ardokhsho on 
the reverse is also seen in a series that has been given the tribal name of Shiladas. These 
coins, which are found in central Panjab, have, on the obverse, the names of Bhodva, Pasan 
and Baeharna, who appear to be rulers of the Shilada tribe holding sway over this area. 

In the next series of coins we find the tribal name of Gadahara, who probably succeeded 
the Shilada rulers in central Panjab. These coins also show the names of Peraya, Kirada 
and Samudra. It is the name Kirada that has also been read in the coins of Kidara 12 (see 
below). Much controversy has been generated by the occurrence of the name Samudra: it 
is generally assumed that it refers to the great Gupta ruler Samudra (320-375) who, in his 
Allahabad pillar inscription, 13 claims to have allowed the daivaputra-shdhi-shdhdnushdhi 
(obviously referring to the Kushan king) to rule in his kingdom and issue coinage in sub- 
ordinate relation to him. 14 It attests to the autonomy of the Kushan king's rule over the 
territory. This statement may be judged against the background of the military successes 
of Samudragupta, who by this time had uprooted (see Chapter 8) the tribal states of the 
Yaudheyas and Madras, whose territories extended over eastern and central Panjab. Thus, 
although Samudragupta had gone to the very border of Gadahara territory, he did not uproot 
that tribe, but merely brought them under his suzerainty. It is equally possible that, being 
pressed by the Sasanians in the west, the Gadaharas established diplomatic ties with the 
Guptas and thus secured themselves a temporary reprieve. 

Before discussing the consequences of this diplomacy, it is necessary to identify the 
Gadaharas. Cunningham 15 long ago suggested that they may be the same people who are 
today known as Gakkharas in western Panjab. Although there is a phonetic similarity in the 
names, the Gakkharas themselves trace their ancestry to the Kayanians in Iran and believe 
that they arrived in the area at the time of Sultan Mahmud of Ghazni (988-1030). On the 
other hand, the Allahabad inscription gives all the Kushan titles. They are taken to apply 
to the Gadaharas, who at this time were issuing Kushan types of coins in central Panjab. 
From this evidence the Gadaharas are taken to be the last of the Eastern Kushans ruling 

11 Zeimal, 1983, p. 225. 

12 Altekar and Majumdar, 1946, p. 105. 

13 Sircar, 1939, p. 258. 

14 Chhabra and Gai, 1981, p. 218. 

15 Cunningham, 1893-94. 



ISBN 978-92-3-103211-0 Contents Eastern Kushans 

independently in Panjab. On the basis of the Allahabad inscription, they continued to rule 
well into the first half of the fourth century. 

The economy of the kingdom can be judged from a variety of sources. First, the con- 
traction of the empire considerably reduced the economic resources of the state. The inter- 
national trade routes that had earlier supplied gold and other luxury items passed out of the 
hands of the Eastern Kushans, a loss that is clearly reflected in the currency of the time. 
The complete cessation of silver currency by the Eastern Kushans, and the debasement of 
the gold currency, show the adverse affect on the balance of trade. However, the abundance 
of copper currency proves the continuation of local demand. The Buddhist monasteries 
and Gandharan art flourished, there was no diminution in the production of works of art 
and urban life continued to be prosperous. The construction of the new city of Sirsukh at 
Taxila and the ruins of Rajar at Charsadda in Gandhara, together with the large number of 
urban settlements, speak of the prosperity of the people. This development must have been 
related to improved agricultural technology. Three types of irrigation project have been 
noted in Gandhara and Panjab. The local governors derived their power from control over 
this irrigation system. While it led to the strengthening of the local authority, it nevertheless 
increased the crops. There is evidence to show that iron mining was undertaken in the Kala 
Chitta range right up to Kalabagh on the Indus, whereas gold was extracted from the 'gold 
ant-hills' in Baltistan and Ladakh. Trade in precious stones continued with China. 

In the land of the Eastern Kushans (i.e. Gandhara and Panjab), no contemporary royal 
inscriptions have been found but many private donations to Buddhist monasteries have 
been recorded in Gandhari Prakrit in KharosthT script. However, the use of Indian Brahmi 
began to spread from this time. Simultaneously, as recorded in the later accounts of the Chi- 
nese pilgrims, the Buddhist monks in the famous Kanishka vihdra (monastery) at Peshawar 
used Sanskrit as their means of expression. The widespread popularity of Shiva and Nandi, 
and other deities such as Karttikeya and Durga, may point to the importance of the Hindu 
religion. On the other hand, Buddhism, which was widespread in this land, developed new 
monastic orders with lavish facilities. The role of the Mahayana, implying the cult of the 
Bodhisattvas, became increasingly prominent among the people. These developments in 
Buddhism may also show the impact of new ethnic elements upon the population of the 
area, although it is impossible to establish the resultant changes in the social system. The 
smrtis (books of sacred law) present a picture of a society that is more applicable to mid- 
India, which was outside the territory of the Eastern Kushans, where Buddhist society con- 
tinued to maintain its hold on the people. The social system is described by the law texts 
as that of vrdtya (land holding), in other words, a system that fell outside the recognized 
rules of the orthodox Hindu system. 

1 70 


ISBN 978-92-3-10321 1-0 Contents Kidarites in Gandhara and Kashmir 

Kidarites in Gandhara and Kashmir 

The earlier history of the Kidarites has been given above (see Chapter 5). The gold coins 
present the Kidarite king standing beside an altar on the obverse and an enthroned goddess 
on the reverse. Under the king's arm the legend Kiddra is written perpendicularly on the 
obverse while the reverse gives the names SrT Shdhi Kiddra, KritavTrya, Sarvayasa, Bha- 
van, Siladitya, Prakasa and Kusala. As the name Kiddra is attached to the term Kushana 
on the coins of Kidara, it is clear that Kidara must have conquered some parts of the 
Kushan territory and hence used the title Kiddra Kushana Shdhi. Other names that appear 
on the coins of Kidara may have been those of his governors posted in the different parts 
of his territory or even of those rulers who succeeded him. If this historical sequence is 
correct, Kidara must have established himself in Gandhara some time in the late fourth 
century. Altekar 16 assumes that Kidara extended his power eastward after the death of 
Samudragupta. Kidara was succeeded by his son, and then by Pira at a later date. 

On the basis of his analysis of the evidence from Kalhana's Rdjatararigini, Harmatta 
concludes that Kimnara 'represents the Kidarite Hun king Kidara whose reign in Kashmir 
must have been presumed already on the basis of historical considerations and numismatic 
evidence but who could not be identified in the text of the Rdjatararigini hitherto' . He gives 
the approximate date of 400-410 for the beginning of Kidara's reign in Kashmir (see pages 
123-4 above). 17 

It should be noted that the volume of gold coins circulated by the Kidarite rulers was 
considerably less than that of the Great Kushans, probably owing to the decline of com- 
merce and the Kidarites' loss of control over the international trade routes. 

With the coming of the Kidarites, Sasanian cultural penetration intensified, and we can 
note the influence of the Zoroastrian religion. This has been well documented by the pres- 
ence of a fire altar on the coins and also by fire worship on the bases of many sculptures 
found in Gandhara. Persepolitan art motifs, such as the Persepolitan bell capital and winged 
animals, continued to be used in Gandhara art. Nevertheless, Buddhism was not eclipsed. 
It was during the rule of the Kidarites that the Chinese pilgrim Fa-hsien (c. 400) visited 
these lands and his account suggests the continuing influence of Buddhism. 

Fa-hsien 18 reached the country of To-li (probably the Darel valley in Gilgit region), 
where he speaks of a society of priests all belonging to the Little Vehicle (the Hinayana 
school). He also describes a wooden image of Maitreya Buddha, 80 feet (24m) high, to 

16 Altekar and Majumdar, 1946, p. 21. 

17 Harmatta, 1984, pp. 185-9. 

18 Beal, 1969, Vol. 1, Introduction, pp. xxix-xxxvii. 



ISBN 978-92-3-10321 1-0 Contents Kidarites in Gandhara and Kashmir 

whom 'the kings of the countries round vie with each other in their religious offerings' . He 
then proceeded towards Wu-Chang (Swat) via the Indus valley road and writes: 

The country of Wu-Chang commences North India. The language of mid-India is used by 
all. Mid-India is what they call the middle country. The dress of the people, their food and 
drink are also the same as in the middle country. The religion of Buddha is very flourishing. 
The places where the priests stop and lodge they call sanghdrdmas. In all there are 500 
sanghdrdmas; they belong to the Little Vehicle without exception. 

Fa-hsien also went to Gandhara, where he describes a great stupa, adorned with silver 
and gold. According to him, the people here also mostly followed the Little Vehicle. Then 
he came to Chu-Ch'a-Shi-lo (the modern Taxila), where great stupas had been built. He 
remarks: 'The kings, ministers and people of the neighbouring countries vie with one 
another in their offerings, scattering flowers and lighting lamps without intermission.' 

The most important description that he provides, however, is of Fu-lousha (identified 
with Peshawar), where he speaks of Kanishka's vihdra and the great stupa. The last he 
calls 'Buddha tower', saying that it is 40 chang and more in height and adorned with all 
the precious substances. Of all the stupas and temples he has seen, none can compare with 
this for beauty of form and strength. This was the highest tower in Jambudvipa. Fa-hsien 
also refers to the alms-bowl of the Buddha (which was still in Peshawar at that time) and 
notes how a king of the Yueh-chih wished to carry it away but, failing to do so, built a 
stupa and a sangharama on the spot. There were 700 priests to look after the bowl and a 
daily ceremony was held in connection with it. At the approach of noon it was brought out 
by the priests and the upasakas (laity) made all kinds of offerings to it before having their 
midday meal. 

Next Fa-hsien went alone to the city of Hi-lo (modern Hadda near Jalalabad in 
Afghanistan), which contained the vihdra of the skull-bone of the Buddha. The ceremony 
is graphically described: 

The door being opened, using scented water they wash their hands and bring out the skull- 
bone of the Buddha. They place it outside the vihdra on a high throne; taking a circular stand 
of the seven precious substances, the stand is placed below it and a glass bell as a cover over 

it. All these are adorned with pearls and gems The offerings finished, each one in order 

puts it on his head (worships it) and departs. Entering by the east door and leaving by the 
west, the king every morning thus offers and worships, after which he attends to state affairs. 
Householders and elder-men also first offer worship and then attend to family affairs. 19 

Fa-hsien then travelled south into the Rohi country before crossing over to Bannu and 
then to Bhira in Panjab. On the way, he 'passed very many temples one after another, 

19 Beat, 1969, Vol. 1, Introduction, p. xxxiv. 



ISBN 978-92-3-103211-0 Contents Later Hephthalites 

with some myriads of priests in them' . At last he arrived at the city of Mo-tu-li (present- 
day Mathura), beyond which lay the middle country (Madhyadesha). Without naming the 
ruler or ruling dynasty of this region, Fa-hsien speaks of the social conditions, the laws of 
punishment, the habits of the people and the position of the chandalas (outcasts of Indian 
society), who are dubbed 'evil men' and hence 'dwell apart from others'. He provides a 
very detailed account of Buddhism in this area and relates a number of traditional stories 
about many cities in this country. From here he went on to Sri Lanka and finally back to 
China by sea. 

Later Hephthalites 

Since the early history of the Hephthalites has been given already (see Chapter 6), we shall 
concentrate here on the later history, with particular reference to Indian sources. In Indian 
works, the Hephthalites are known as Sveta Huna; they are designated by these names 
in the Brhatsamhitd of Varahamihira and in the Mahdbhdrata. 20 On the evidence of the 
Avesta, Bailey has identified Sveta Huna with Spet Hydn (meaning White Huns), and Hara 
Huna with Karmir Hydn (meaning Red Huns). 21 Thus the Huns were split into two groups 
- Red Huns and White Huns - in the Indian literature. The struggle of the White Huns 
against the Gupta emperors, and their establishment of an independent empire south of the 
Hindu Kush, has been discussed earlier (see Chapter 6). Harmatta's view that there was a 
Khingil dynasty ruling over Kabul must now be modified in the light of a fresh interpre- 
tation of the source material from Kalhana's RajataranginT, coins and inscriptions. 22 The 
opinion given below follows the reconstruction made by Biswas, 23 who has listed a num- 
ber of Huna kings from the Puranas, and from the RajataranginT. The most likely person 
to have succeeded Mihirakula in Kashmir and Gandhara is Pravarasena II, who is believed 
to be the son of Toramana. However, there was an interregnum between these two rulers. 
Pravarasena came to the throne soon after 530 and ruled for about 60 years. He founded a 
city with his own name, Pravarasenapura (identified with modern Srinagar), and adorned 
it with markets. Here he also built the great temple of Pravaresha. Within the city he con- 
structed a causeway or bridge. His coins bear, on the obverse, the figure of a standing king 
and two figures seated below right and left, with his name Pravarasena; and on the reverse, 
a goddess seated on a lion with the legend Kidara. The significance of this legend is not 
known. This king was followed by Gokarna, some of whose coins have been discovered. 

20 Biswas, 1973, pp. 26-8, where quotations from the original are given. 

21 Bailey, 1954, pp. 12-16; 1932, p. 945. 

22 Harmatta, 1969, p. 404. 

23 Biswas, 1973, Chs. 5 and 6. 



ISBN 978-92-3-103211-0 Contents Later Hephthalites 

He established the shrine of Shiva, called Gokarnesvara. His son Narendraditya, who bore 
the second name of Khihkhila, also consecrated shrines to Shiva, called Bhutesvara. His 
son was Yudhishthira, nicknamed Andha-Yudhishthira on account of his small eyes. 24 

Narendraditya Khihkhila is identified with a king whose name appears at the base of the 
stone image of Vinayaka (Ganesha), found in Kabul, but probably from Gardez. The king's 
name is recorded as Parama-bhattaraka Maharajadhiraja ShrT Shahi Khingala Odya (tya) 
na-Shdhi. He also issued coins with the name either Deva Shahi Khingila or Shri Narendra. 
Some coins have the legend Kiddra under the king's arm. All these kings are identified as 
one and the same by Biswas, who maintains that Khihkhila ruled a domain stretching from 
Kashmir to Kabul. According to Kalhana, he ruled for some 30-36 years, that is roughly 
between 597 and 633. According to Biswas, the Chinese Buddhist pilgrim Hsiian-tsang 
came to Kashmir when Khihkhila was ruling here. Regarding the extent of his empire, 
Biswas concludes: 

The empire of Kashmir included the Kabul valley, the Swat valley and the mountain regions 
of Kashmir proper and in the south-east extended as far as Sakala on the Chenab river. If the 
king of Kashmir had a hold over Swat, the Kabul valley and Bannu, it is possible that his 
empire extended even to Gardez. The Gardez inscription of Khingala was probably thus of 
the Kashmir king Khinkhila, who was also the overlord of Udyana. 25 

Khihkhila was succeeded by his son Yudhishthira who, according to the Rdjatarariginl, 
ruled for 40 years (until c. 670), when he was dethroned by Pratapaditya, son of Durlab- 
havardhana, the ruler of the Karkota dynasty. According to another version Yudhishthira 
ruled for only 24 years, or until c. 657. Although Yudishthira was the last great independent 
Huna (White Hun) ruler, Kalhana gives a further line of his successors who continued to 
rule in subordinate positions in Kashmir and other areas. 

The end of the rule of Yudhishthira brought further changes in the Huna kingdom. One 
major consequence was the foundation of the so-called Turk Shahi dynasty in Kabul and 
Gandhara, whose history and origins have been reconstructed in great detail by Rahman: 26 

The history of this Turkish family can be traced back to at least a.d. 666, when a Rutbil is for 
the first time mentioned in the Arabic chronicles. The date of Barhatigin who, according to 
Albiruni [al-Biruni], was the founder of the Turk Shahi dynasty, must therefore fall about a.d. 
666. It would seem that Barhatigin and the first Rutbil were brothers. The dynastic change 
mentioned by Huei Ch'ao [Huei-ch'ao] appears to have taken place long before his visit, but 
he came to know of it only when he was in Gandhara in a.d. 726 and at that time he mentioned 
it in the account of his journey. Thus the date of the beginning of the rule of the Turk Shahis 

24 Stein, 1900-1, Book 1, verses 346-50. 

25 Biswas, 1973, p. 137. 

26 Rahman, 1979, pp. 37-47 and Ch. 4. 



ISBN 978-92-3-103211-0 Contents Later Hephthalites 

may be placed around a.d. 666 or slightly earlier 27 [i.e. immediately after the overthrow of 
Yudhishthira by the Karkota dynasty in Kashmir] . 

The Turk Shahis remained in power for nearly 177 years. The end of their rule in Kabul is 
dated to 843 on the basis of epigraphic evidence. But the western branch of the Turk Shahis 
(the Rutbils of Arabic sources, and generally known as rulers of Rukhkhaj) continued to 
rule a little longer until the rise of Ya c qubb. Laith, the amir of Seistan (modern Sistan), 
who captured Kabul in 870. The last of the Rutbils, a fugitive in Kabul or Zabulistan, was 
captured in 870 - a date which finally brought to a close the long history of the Turkic- 
speaking Hephthalites. 

Rahman rightly points out that the political history of the Turk Shahis is inextricably 
linked with the history of the Muslim governors of Seistan. How Barhatakin came to power 
is wrapped in mystery, although it is probable that his base was in Gandhara. Sheltered 
behind the rugged hills of the Khyber Pass, he built up his strength slowly and waited for 
his chance when the kingdom of Kabul and Zabulistan were subject to repeated attacks 
by the Arab governors of Seistan. In the wake of two attacks by Ibn Samura, Barhatakin 
gathered his forces and attacked Kabul. We learn from Chinese sources that the ruler of 
Kapisa (probably Khihkhila) was killed and Barhatakin proclaimed himself king of Kabul. 
He extended his rule to Zabulistan and appointed his brother as Hindu governor with the 
title of Rutbil, i.e. 'war thruster'. This is sometimes corrected to hitivira, 2 ^ but Harmatta 
takes it for Zubil and connects it with yabghu. 29 

From these sources 30 Harmatta builds up the following chronology of the Turk Shahi 

Wu-san T'e-chin Shai, 720-738 (Shri Tagino Shaho on the coins) 
Fu-lin-chi-so, 738-745 (Phromo Kesaro on the coins) 
Po-fu-chun, 745 onwards 
Ju-lo-li in Gandhara, 759-764. 

From a Tibetan source, Harmatta 31 quotes the name Phrom Ge-sar and identifies him with 
Fu-lin-chi-so, mentioned above. Further, on the basis of his interpretation of the Tochi 
valley inscriptions, Harmatta 32 concludes that the Sanskrit inscription there mentions the 

27 Rahman, 1979, p. 47. 

28 Ibid., p. 180. 

29 Harmatta, 1969, p. 406. 

30 Ibid., p. 409. 

31 Ibid., pp. 409-11. 

32 Ibid., p. 367. 



ISBN 978-92-3-103211-0 Contents Economic and cultural progress 

name of (Mihira) Bhoja of the Gurjara Pratihara dynasty who, in 860, extended his rule 
westward and helped Lalliya, the founder of the Hindu Shahi dynasty, to wrest Gandhara 
and Kabul from the Turk Shahis and kept him there in opposition to Ya c qub b. Laith. This 
inference of Harmatta brings the Gurjara dominion right beyond the Indus, for which we 
have no other evidence. However, the power of the Gurjaras certainly increased in eastern 
and central Panjab and the Karkota dynasty ruled supreme in Kashmir. As a result of their 
aggrandizement, Huna power collapsed in the third quarter of the ninth century; those Huna 
principalities that survived became assimilated to the local order and thereafter played an 
insignificant role. 

Economic and cultural progress 

The archaeological evidence from Taxila led Marshall 33 to speak of the great destruction 
caused by the Huns and the consequent disruption of the economic and cultural progress 
of the countries where they ruled. This conclusion has been contradicted by Dani, 34 who 
believes that urban life continued in Taxila and the monasteries were maintained, as attested 
by Hsiian-tsang during his visit in the seventh century. New evidence from along the 
Karakorum highway reveals a brisk trade and commercial relations between Gandhara, 
China and the trans-Pamir region. Although the Silk Route was disrupted because of new 
imperial alignments, trade was deflected southward. 

One major change in this period relates to agricultural production and the administration 
of revenues. Up to the time of the Kidarites, there is evidence for the survival of the satrapal 
system of administration in the Kabul valley, Gandhara and Panjab; but during their rule 
this administrative system appears to have died out. In its place we note a large number of 
tribal chiefs who assumed the title of raja. For the first time Bana (court poet from 606 to 
647) in his Harsa-Caritam uses the title of rajaputra, from which is derived the modern 
term Rajput. The growth of the rajas, and Rajputs (see Chapter 8) in the socio-economic 
life of the hilly regions and plains of Panjab is a new phenomenon that dates from the time 
of the Huns. The changing pattern of land tenure led to a new form of economic system, 
which has been loosely described as a feudal relation, although feudalism of the European 
pattern did not develop in this part of the world (see also Chapter 8). The Rajput system 
perpetuated the claim of the tribal heads to land which they possessed by right of their 
joint aggrandizement. Thus they became the real owners of the land, and also of the settlers 
upon it. 

33 Marshall, 1951. 

34 Dani, 1986, pp. 75-8 



ISBN 978-92-3-103211-0 Contents Economic and cultural progress 

This new property system led to the development of a social order that has survived to 
the present time. Whether it also led to greater agricultural production is difficult to say. 
It undoubtedly led to new agricultural management by the officers of the state through the 
tribal heads who had a direct stake in the land. There are at least three pieces of archae- 
ological evidence: from the Idak-Spinwam region in north Waziristan, from Gilgit proper 
and from Skardu. In all these places new irrigation channels were opened up. In other 
areas, natural springs were channelled to irrigate terraced fields. Consequently, there does 
not appear to have been any loss in agricultural production although the landless labourers 
undoubtedly suffered and slavery must have been rampant as a consequence. For example, 
the Turk Shahi ruler of Kabul had to pay an annual tribute of 2,000 slaves to the Arab 
governor of Khurasan. 

On the other hand, the system led to relations of production in which the agricultural 
magnates enjoyed all the economic, social and even religious privileges whereas ordinary 
people struggled to survive and were at the mercy of the landlords. As amply demonstrated 
by the history of the rulers of Gilgit, there were continuous wars of succession between 
the sons of these chiefs, and a consequent wastage of property and manpower. The feu- 
dal Rajput system nevertheless established itself in the existing social milieu. Rather than 
destroying the caste system, it found its place within it, absorbing the caste groups within 
its own economic sphere and giving them a new function. 

The Huns were fervent worshippers of the sun god and of Shiva and a number of Shiva 
temples were built in Kashmir. In the Gilgit region, Buddhism flourished and developed 
a new form. The most important piece of evidence comes from the Buddhist creations at 
Bamiyan, where tall Buddha figures, cave paintings and monasteries attest the progress of 
art in this region (see Chapter 6). 

Hsiian-tsang has left a detailed description of the Buddhist centres and monastic life in 
the period of the Huns, 35 waxing lyrical when he visits Bamiyan. From a cultural point of 
view, his most valuable observation is the following: 

These people are remarkable, among all their neighbours, for a love of religion (a heart of 
pure faith); from the highest form of worship to the three jewels, down to the worship of 
the hundred {i.e. different) spirits, there is not the least absence {decrease) of earnestness and 
the utmost devotion of heart. The merchants, in arranging their prices as they come and go, 
fall in with the signs afforded by the spirits. If good, they act accordingly; if evil, they seek 
to propitiate the powers. There are ten convents and about 1000 priests. They belong to the 
Little Vehicle, and the school of the Lokottaravadins. 36 

The description of Kapisa is no less instructive in its picture of the economy and culture: 

35 Beat, 1969, Vol. 1, p. 50. 

36 Ibid. 



ISBN 978-92-3-103211-0 Contents Sanskrit references 

It produces cereals of all sorts, and many kinds of fruit-trees. The shen horses are bred here, 
and there is also the scent (scented root) called Yu-kin. Here also are found objects of mer- 
chandise from all parts. . . In commerce they use gold and silver coins, and also little copper 
coins. . . The king is a Kshattriya by caste. He is of a shrewd character (nature), and being 
brave and determined, he has brought into subjection the neighbouring countries, some ten of 
which he rules. 37 

Hsiian-tsang's descriptions of the capital cities of Kapisa, Gandhara and Taxila leave no 
doubt that these centres continued to maintain their urban nature in this period, although 
some were no longer royal seats of government. On the other hand, the foundation of 
new cities in Kashmir by the later Huna kings, as noted by Kalhana, speaks highly of the 
patronage they exercised. Under their rule Shaivism and the worship of the sun god were 
encouraged (many images of the sun god have been found in Gandhara). But as far as the 
old cities such as Taxila and Purushapura are concerned, fresh archaeological evidence has 
not produced any new data. Only in the case of Taxila do new studies of the earlier finds 
suggest that the fortifications at the site of Giri belonged to the Huna period. On the other 
hand, Huei-ch'ao's visit to Purushapura in 726 and his description of the Kanishka vihdra 
there provide ample proof of the continued existence of the Buddhist centre. 

Sanskrit references 

The references in Sanskrit sources to the dynasty of Mihirakula are of great importance as 
they throw light on the character of the rulers. The largest number of references is found in 
Kalhana's Rdjatarangirii?^ In the Purdhas, the Hunas are equated with the Mlecchas and 
are said to rule over the vrdtya countries (see above). The Prakrit work, the Kuvalayamdld, 
mentions the land of Uttarapatha through which flows the River Chandrabhaga (Chenab). 
On its bank lay the city of Pavvaiya, where lived Shri Toramana (the father of Mihirakula), 
enjoying the sovereignty of the world. In the Kura inscription, 39 which records the con- 
struction of a Buddhist monastery, the ruling king is thus recorded: Rdjadhirdja Maharaja 
Toramana Shdhi Jaula. The Gwalior inscription 40 paints a memorable picture: 

[There was] a ruler of [the earth] of great merit, who was renowned by the name of Sri 
Toramana, by whom, through his heroism [that was especially characterized by] truthfulness, 
the earth was governed with justice. Of him, the fame of whose family has risen high, the son 

37 Beal, 1969, Vol. 1, pp. 54-5. 

38 On the basis of this, Biswas has reconstructed the political history of the Hunas in India; see Biswas, 

39 Sircar, 1939, p. 56. 

40 Ibid., p. 400. 



ISBN 978-92-3-103211-0 Contents Coinage of the Hunas 

[is] of unequalled power, the lord of the earth, who is renowned under the name of Mihirakula 
[and] who [himself] unbroken [?] worships Pasupati. 

These quotations show the nature of the Huna rulers who conquered this part of the world. 
Mihirakula was a devotee of Shiva. That he wielded great power is confirmed by the Man- 
dasor inscription of Yashodharman, which says of Mihirakula: 'Through the embraces of 
whose arm Himalaya carries no longer the pride of the title of being [an inaccessible] 
fortress.' The Bhitari pillar inscription of Skandagupta describes the eventful scene of the 
Gupta king's terrible conflict with the Hunas in the following words: 'By whose two arms 
the earth was shaken, when he, the creator of a terrible whirlpool, joined in conflict with 
the Hunas.' 

Bana, court poet of King Harsha, speaking of Harsha's father in the early seventh cen- 
tury, uses the phrase Huna-Harina-Kesari (lion to the Huna deer). In other words, from 
the great power that the Hunas wielded in their early career and so graphically described 
in Sanskrit literature, they lost their prestige after their defeat in c. 528 by Yashodharman 
(king of Malwa) and Baladitya (the Gupta king of Magadha), and were remembered as 
weak as deer before the lion king of Kanauj. 

Coinage of the Hunas 

The earliest Huna coins imitate those of Shapur II, except that the Pahlavi script is replaced 
by Bactrian. These coins also bear the Hephthalite symbol on the obverse. They do not bear 
mint marks, although a few coins show simple BrahmT letters, such as Thai, Sa, or Se, Je, 
Bra, Tu or Dhe. These Huna coins are divided into three groups: (a) Early Huna coins; (b) 
Huna coins of Tunjina I, Toramana and Mihirakula; and (c) Later Huna coins of kings who 
ruled in Kashmir, Gandhara and parts of Panjab. 

The BrahmT legends of the second group of Huna coins establish their identity and 
also show more varieties. One coin of the earlier design has yielded the name Thujana 
(*Thumjina) and is attributed to the first ruler, Tunjina. Then we find coins inscribed 
Shahi Javukha or Shahi Javuvla. The attribution of these coins to Toramana is doubt- 
ful. His coins are only in silver and copper: no gold coins of his time have so far been 
found. Toramana's silver coins are of three varieties: those showing the Sasanian bust 
and fire altar; those showing a horseman on the obverse; and those copying the Gupta 
coins from Gujarat. The reverse of the last series depicts a dancing peacock. The coins 
bear the BrahmT legend meaning 'Shri Toramana Deva, the invincible, conquers'. His cop- 
per coins are of two categories: the first shows the Kushan type of standing king on the 
obverse with the king's name and a seated goddess on the reverse; the second variety 

1 79 



has a Sasanian bust on the obverse, and a solar wheel on the reverse with Tora in bold 

Mihirakula struck coins in many styles. They are also of silver and copper. The silver 
coins show a bull on the obverse with the legend Jayatu Mihirakula or Jayatu Vrsadhvaja. 
In examples where the bull is seated, the king has his face turned to the right; on the reverse 
is the fire altar. Mihirakula's copper coins are more common: they keep the bust type of 
the silver coins but omit the bull standard. On the reverse, however, the bull is shown in 
the upper register. Sometimes there is a trident before the bull with the legend Jayatu Vrsa. 
The second variety copies the Kushan example of standing king on the obverse with the 
legend Shahi Mihirakula. On the reverse again we find the bull. 

The Later Huna coins of Pravarasena II (which are gold) continue the Kushan style of 
the standing king on the obverse and the seated goddess on the reverse, with the legend 
Shri Pravarasena on the obverse and Kidara on the reverse. The coins of Narendraditya 
Khihkhila are struck with his name, Deva Shahi Khihgila. These coins have the beardless 
king's head to the right. Another type shows the standing king on the obverse with the 
legend Kidara, and the seated goddess with the legend Shri Narendra on the reverse. We 
also find the coins of Lahkhana Udayaditya showing the bust of the king. Thus the coins 
of the Hunas, which begin by copying the Sasanian type, later show the Kushan type and 
then gradually become more localized with Brahmi legends and symbols. 41 

Part Two 


(B. A. Litvinsky and M. H. Zamir Safi) 

Between 560 and 563, the Turks inflicted a crushing blow on the Hephthalites. The 
fact that Sughd resumed its external political relations in 564, after a long interval, enables 
us to pinpoint this date more precisely: Sughd could no longer have been controlled by 
the Hephthalites, and the decisive battle against the Turks at Bukhara probably took place 
in 563. According to Firdausi's Shdh-ndme, troops from Balkh, Shughnan, Amol, Zamm, 

41 For another classification, chronological consequences and an analysis of coins, inscriptions, symbols, 
etc. of Huna coins, see the very important work of Gobi, 1967, Vols. 1-4. 




Khuttal, Termez and Washgird fought on the side of the Hephthalites in this battle and 
weapons and essential equipment were also obtained from those places, 42 that is chiefly 
from the right-bank regions of Tokharistan (ancient Bactria). 

After the battle, the remnants of the Hephthalite levies fled to the south, where Faganish, 
the ruler of Chaganiyan, was chosen as king. Upon learning of these events, the shahanshah 
of Iran, Khusrau I (531-579), moved his troops, obliging Faganish to accept vassal status. 
The areas to the north of the Amu Darya (Oxus) were subsequently recognized as posses- 
sions of the Turk kaghans while the areas to the south were acknowledged as belonging 
to Sasanian Iran. It is reported that the Turks exacted tribute from the Hephthalites (evi- 
dently, the Hephthalites of northern Tokharistan). It is also recorded that Khotan, Persia 
and the Hephthalites (i.e. the same Hephthalites of Tokharistan) rebelled against the Turks 
in581. 43 

Al-Biruni writes in his al-Qanun al-Mas c udl that Tokharistan 'in the days of old was 
the country of the al-Hayatila [Hephthalites]'. 44 According to modern researchers, the 
Islamic geographic term Haital (Hephthalite) 'was for long synonymous with the regions 
of Tuxaristan and Badaxsan to the south of the upper Oxus and those of Chaganian, 
Qubadiyan, Xuttal and Waxs to the north of it'. 45 

Thus Hephthalite buffer principalities with vassal status were formed in the south of 
Central Asia. One of them, Chaganiyan, lies in the upper and central valley of the Surkhan 
Darya river. It is certain that a Hephthalite dynasty - which may have been descended 
from the Faganish mentioned above - ruled in Chaganiyan. The coinage in circulation was 
mainly that of Khusrau I Anushirvan: at first, this was the genuine currency of the shahan- 
shah, but imitations later appeared with the name of the local rulers, 'oappo x<5>?o, %apivo 
xSno' . Finally, coins appeared stamped like those of Khusrau I but with the name of the 
local ruler, ' jtolvolo x& r l \ on me reverse, on either side of an altar; while the obverse 
bears no inscription. 46 The local ruling dynasty, whose representatives bore the title of 
Chaghan khudat, continued to exist in the pre-Arab period. 47 Several of the rulers are 
known to us by name: an Afrasiab inscription states that emissaries arrived in Samarkand 
from the Chaganiyan ruler, Turantash. Later, in the first quarter of the eighth century, the 
ruler of Chaganiyan was Tish 48 the 'One-Eyed', who also ruled the whole of Tokharis- 
tan with the title of yabghu. There was also a developed system of administration. The 

42 Mohl, 1868, pp. 308-16. 

43 For further details, see Mandel'shtam, 1964, pp. 42-3. 

44 See al-Biruni, 1973, p. 467. 

45 Bosworth and Clauson, 1965, p. 5. 

46 Rtveladze, 1983, p. 75. 

47 Bosworth, 1981, pp. 1-2. 

48 'Tish' was the Bactrian name for the star Sirius. 




above-mentioned emissaries from Chaganiyan were led by the dapirpat, the chief scribe or 
head of chancellery. 

Another major Hephthalite possession was Khuttal, which lay within the territory of 
the present-day Kulyab region, that is, the basin of the River Kyzyl-su, and at times also 
included the Vakhsh valley. The local dynasty here also followed an established order of 
succession. 49 The local rulers bore the Iranian title of khuttal-shah or sher-i khuttal while 
the Arabs referred to them as muluk (pi. of mdlik, king). 50 There is as yet no evidence that 
specifically Khuttal coinage was minted, but the practice existed in several neighbouring 
territories. In particular, the territories of Termez and Kobadian minted their own copper 
coinage from the end of the fifth to the beginning of the seventh century. Pierced copper 
coins bearing cursive Hephthalite inscriptions circulated in Kobadian and Vakhsh from the 
second quarter of the seventh century. 51 

A further important Hephthalite possession was Balkh, the premier town in Tokharistan, 
which, at that time, had extensive territory. Written reports provide detailed descriptions of 
Balkh and its buildings, including the renowned Naubahar (Buddhist temple near Balkh) 
of the late sixth century. Its name derives from the Sanskrit nova vihara (new monastery). 
The Buddhist community was headed by the barmak, a title derived from the Sanskrit, 
parmak (superior or chief). At Balkh, too, coins were minted. 

Important information about all these territories is given in the description of the journey 
made by the Chinese Buddhist pilgrim, Hsiian-tsang; although he travelled after Tokharis- 
tan had been conquered by the Turks, much remained unchanged at the time of his journey 
in 630. Tokharistan (Tou-ho-lo) comprised 27 territories 'divided by natural boundaries'. 
Of the local population he says: 'Their language differs somewhat from that of other coun- 
tries.' His description of the written language corresponds to Bactrian Hephthalite writing, 
which was based on the Greek alphabet. Hsiian-tsang also mentions the wealth of literary 
works and remarks that 'most of the people use fine cotton for their dress; some use wool' . 
He refers to their coinage, which differs from that of other countries. 

The territory of Termez (Ta-mi) lay on an east-west axis, as did its capital, which con- 
tained some 10 Buddhist sarigharamas (monasteries) with approximately 1,000 monks, 
and stupas and images of the Buddha. The territory of Chaganiyan (Shih-han-na) was 
somewhat smaller than Termez, its capital was only half the size and there were only some 
500 sangharamas. Similarly, the territory and capital of Kobadian (Kio-ho-yen-na) were 
half the size of those of Termez and the country contained 10 monasteries with hundreds 

49 Noldeke, Tabari, 1973, Vol. 2, p. 1618. 

50 Marquart, 1901, p. 30; Belenitskiy, 1950, p. 115. 

51 Davidovich and Zeimal, 1980, pp. 72-4. 




of monks. Although the territory of Vakhsh was slightly smaller than that of Termez, their 
capitals were almost the same size. The capital of the territory of Khuttal (Kho-to-lo) was 
the same size as Termez. 

The territory of Balkh (Po-ho) was larger than that of Termez and bordered on the River 
Amu Darya to the north. The capital was the same size as Termez, well fortified but with a 
small population. Agricultural produce was varied. There were roughly 100 sahghdrdmas 
with 3,000 Hinayana monks. Outside the town was the 'new sanghardma' , 'which was 
built by a former king of this country' . There follows a description of this sanghardma and 
the buildings in the religious complex, including a giant stupa. 52 

The Hephthalites settled over a much wider area within the limits of modern Afghanistan 
than the area of Balkh, penetrating westwards as far as Herat and Badhghis. In the strug- 
gle against the Arabs, the tribes of the Herat Hephthalites helped to resist the troops of 
c Abdallah b. Amir in Kuhistan (see Chapter 19, Part One). 53 The Hephthalites are men- 
tioned in connection with the events of 704, 54 along with 'Tibetans' and ' Turks'. The 
leader of the Hephthalites of Herat and Badhghis, Tarkhan Nizak (Figs. 1 and 2), played 
a major role in the struggle against the Arabs. Arab sources provide detailed informa- 
tion about this ruler and his role in the events connected with the fall of the last Sasanian 
king, Yazdgird III (632-651). According to the early thirteenthcentury geographer Yakut 
(V, 461), Badhghis was the 'headquarters of the Hayatila [Hephthalites]'. Other sources 
describe Tarkhan Nizak as 'king of the Hephthalites'. He also played an active part in 
the struggle against the Arabs in Tokharistan and twice attempted to capture Balkh. Taken 

FIG. 1. Coins of Tarkhan Nizak (observe). Photo: © Bibliotheque Nationale de France 

52 Beat, 1969, Vol. 1, pp. 37-46. 

53 Noldeke, Tabari, 1973, Vol. 1, p. 2886 


Ibid., Vol. 2, p. 1,153. 



ISBN 978-92-3-103211-0 Contents the successors of the Hephthalites 

FIG. 2. Coin of Tarkhan Nizak (reverse). Photo: © Bibliotheque Nationale de France 

together, this evidence indicates that a powerful confederation of Hephthalite tribes existed 
in north-western Afghanistan. 55 

The Khalaj, the successors of the Hephthalites 

The Hephthalites were succeeded by the Khalaj, a people or tribes originally living in west- 
ern Turkestan and then in Afghanistan during the ninth (eighth?) to the twelfth centuries. 
Arab geographers of the ninth and tenth centuries place them among the Turk tribes and 
frequently confuse the Khalaj with the Khallukh (i.e. Karluk) as only diacritical marks 
distinguish these two ethnonyms in Arabic script. Hence, information relating to the Khal- 
lukh is often included in descriptions of the Khalaj. For example, the Arab geographer Ibn 
Khurradadhbih includes the Khalaj among the Turk tribes and locates their winter quar- 
ters in the region of the River Talas adjoining the land of the Khallukh, but also states 
that they live 'on this side' of the Amu Darya, i.e. to its south and west. According to the 
tenth-century geographers al-Istakhri and Ibn Hauqal, the Khalaj lived in Zamin-Dawar. 
They are said to have moved in ancient times to that province situated between Hind and 
Seistan (modern Sistan) but to have retained 'Turkic appearance, dress and language'. The 
Persian geography, the Hudud al- c alam [Regions of the World], provides the following 

In Ghazni and in the limits of the boroughs which we have enumerated, live the Khalaj Turks 
who possess many sheep. They wander along climates, grazing grounds and pasture-lands. 

55 Marquart, 1901, pp. 76-8; Markwart, 1938, pp. 39-41; Bivar, 1971, p. 304. 



ISBN 978-92-3-103211-0 Contents the successors of the Hephthalites 

These Khalaj Turks [are] also numerous in the provinces of Balkh, Tukharistan, Bust and 
Guzganan. 56 

A later author (beginning of the thirteenth century), Muhammad b. Najib Bakran, writes in 
his Jahan-name that: 

the Khalaj are a tribe of Turks who from the Khallukh limits emigrated to Zabulistan. Among 
the districts of Ghazni there is a steppe where they reside. Then on account of the heat of 
[the] air their complexion has changed and tended towards blackness; the language, too, has 
undergone alterations and become a different dialect. A tribe of this group went to the limits 
of Bavard and founded some settlements. 57 

The author notes acutely that the appearance and language of the Khalaj differ signifi- 
cantly from those of the Turks. Another source, the Tdrikh-e Sistdn [The History of Sis- 
tan], distinguishes the Khalaj from the Turks when describing the peoples conquered by 
Ya c qub b. Laith. 58 They are also said to differ in Firdausi's Shah-name. 59 The account 
by al-Khwarazmi in his Mafatlh al- c ulum (composed shortly after 977) is conclusive with 
regard to the origin of the Khalaj. He writes that the Hayatila (Hephthalites) 'are a tribal 
group who were formerly powerful and ruled over Tukharistan; the Khaladj and Kandjina. 
Turks are remnants of them.' 60 According to Minorsky, 'the early history of the Khalaj 
tribe is obscure'. 61 

Marquart 62 considers that the Khalaj belonged to the Hephthalite people or Hephthalite 
confederation and that they were a Turkic people. 63 He arrives at the conclusion that they 
belonged to the Hephthalites on the basis of al-Khwarazmi's Mafatlh al- c ulum and by 
analysing 'two names found in pre-Islamic sources' . The first name, Xwls (or Khwls) (Mar- 
quart suggests the reading: *Kholas), occurs in a Syriac history known as the Zaharias 
Rhetor (554-5), which lists the names of the Northern Barbarians. The second name, Kho- 
liatai or Choliatai, occurs three times in the account by Zemarkhos, the Byzantine envoy 
to the Turkic court in 568. For Minorsky, it is clear that the Kholiatai: '(1) lived to the east 
of the Jaxartes, (2) probably to the west of Talas, (3) that they had towns, and (4) that their 
ruler was an important vassal of Dizabul [king of the Turks]'. 64 

56 Hududal- C ulum, 1970, p. 111. 

57 Ibid., p. 348. 

58 Tarikh-e Sistdn, 1976, p. 170. 

59 Mohl, 1868, p. 682. 

60 Bosworth and Clauson, 1965, p. 6. 

61 Minorsky, 1940, p. 426. 

62 Marquart, 1901, pp. 251-4. 

63 Elsewhere he is inclined to believe that their language was Mongolic: see Marquart, 1914, p. 73; Mark- 
wart, 1938, p. 93. 

64 Minorsky, 1940, p. 427. 



ISBN 978-92-3-10321 1-0 Contents Urban life and art in Tokharistan 

In his list of the 'lands of the Turks', Ibn Khurradadhbih mentions the Karluk and the 
Khalaj together 'and these [latter] are on this side of the River [Oxus]', that is, to the south 
and west of the Amu Darya. In another passage, he reports that the winter quarters of 
the Karluk were near Kasra-bash to the south of the Talas, 'and near them are the winter 
quarters of the Khalaj'. But the distance between the Amu Darya and the Talas is such 
that it would have been impossible for the tribes living beyond the Amu Darya to use the 
Talas pastures as winter quarters. The logical conclusion is: 'Either the text is mutilated or 
there were still (?) some Khalaj living near the Khallukh.' 65 Minorsky adds, 'The tempting 
point in Marquart's theory is that both Zemarkhos and Ibn Khurradadhbih have in view the 
region near Talas. However, the identity of the names Xwls xokiarai and Khalaj is still to 
be proved.' 66 

In spite of this and other objections, Marquart's view has become the prevailing one in 
the field. It may be taken as established that the Khalaj were the descendants of the Hep- 
hthalites who moved to the south of Afghanistan, although some of them remained in the 
north. Subsequently, they are frequently referred to in historical accounts as participating 
in various wars. Some of them moved to western Iran (Khalajistan) and even to Anatolia; 
they now speak a very archaic Turkic language. After a number of ethnic transformations, 
the Afghan Khalaj became 'the Pashto- speaking Ghalzay or Ghilzay tribe of Afghans'. 67 
This Pashto-speaking tribe is first referred to in connection with Babur's campaign against 
Ghalji near Ghazni, which means that this process was completed by the beginning of the 
sixteenth century. 68 

Urban life and art in Tokharistan 

The sixth century was characterized by a degree of progress in urban life in Tokharistan. 
Old towns and settlements, including such major ones as Balkh, Dilberjin and Termez, 
continued to exist. There are over 100 early medieval monuments on the territory of the 
Surkhan Darya region of Uzbekistan alone, of which roughly 60 per cent date from the 
fifth to the eighth century. The network of urban settlements was reorganized, perhaps fun- 
damentally, with the development of new social and economic conditions and the begin- 
nings of feudal relationships. Many new settlements sprang up, including some which were 
medium-sized or large. As far as we can judge from the incomplete data, the internal struc- 
ture of urban settlements underwent modification. Despite the far-reaching changes in all 

65 Ibid., p. 428. 

66 Ibid. 

67 Minorsky, 1940; Bosworth and Doerfer, 1978; Rahman, 1979. 

68 Frye, 1965, p. 1001. 



ISBN 978-92-3-10321 1-0 Contents Urban life and art in Tokharistan 

areas of urban life, material culture and art, however, links may be observed with the pre- 
vious period. 

Among the urban centres which finally took shape in the sixth century was Kafyr-kala, 
capital of the Vakhsh valley. The town and citadel formed a regular square with sides of 360 
m. They were surrounded on all sides by a large ditch (50-60 m wide and 5 m deep) and 
by defensive walls reinforced with strong towers. The citadel (of 70 x70 m) in the north- 
east corner of the town had exceptionally strong fortifications. It was surrounded by two 
walls. The inner, main wall had strong angle towers and projecting semi-circular towers. 
Between the towers were stepped arched niches containing false arrowshaped loopholes. 
The fortifications were exceptionally strong. Passages running along the fortified walls 
were used for defensive purposes. The palace complex in the citadel was built around 
a rectangular hall with an area of 200 sq. m, surrounded by smaller halls and domestic 
offices. The southern part of the palace contained the Buddhist vihdra and a courtyard. The 
vihdra had a central sanctuary and ambulatory. The walls of the sanctuary were decorated 
with polychrome murals depicting the Buddha and other Buddhist figures. 

There were also castles, one example being Kuev-kurgan. This was a twostorey struc- 
ture (18x20 m) erected on a 3-m platform of pakhsa (sun-baked brick). The rooms on the 
lower floor were built around a small, almost square hall. There were also several rectan- 
gular and passage-like rooms, some of which were very elongated. The layout of the upper 
floor is uncertain, although a richly decorated ceremonial room was apparently located on 
the upper floor above the hall. This room contained a frieze of some ten or twelve painted 
statues and the walls were covered by murals. 69 

Balalyk-tepe with its remarkable cycle of paintings, the later areas of Dilberjin with 
the corresponding paintings, Bamiyan and a number of other monuments in Afghanistan 
apparently belong to the sixth century, when architecture, painting and decorative metal- 
work achieved a high level of development. A detailed description of the surviving exam- 
ples of the art of the period is provided in Chapter 6. 

69 Litvinsky and Solov'ev, 1985. 



ISBN 978-92-3-10321 1-0 Contents Origin and political history of the Guptas 


K. Chakrabarti 


Origin and political history of the Guptas 188 

Social and economic conditions 193 

Administration 197 

Religious life 199 

Literature 201 

Science 202 

Art and architecture 203 

Origin and political history of the Guptas 

The political disintegration which followed the dissolution of the Kushan Empire continued 
up to the beginning of the fourth century. The Kushans still ruled over western Panjab, but 
they had ceased to exercise any authority further east. The Sakas ruled over Gujarat and 
a part of Malwa, but their power was also on the decline. The rest of northern India was 
divided into a number of small kingdoms and autonomous states. 

The origin of the Guptas is somewhat obscure. Many authorities on Gupta history 
believe that they came from Magadha or northern Bengal, which was the original nucleus 
of their empire. On the basis of the provenance of early Gupta coin hoards and the distri- 
bution of the important Gupta inscriptions, historians have now come to accept the lower 
Doab region as the original home of the Guptas. 

See Map 4. 



ISBN 978-92-3-10321 1-0 Contents Origin and political history of the Guptas 

From the Allahabad pillar inscription of Samudragupta we learn that while the first two 
kings of the Gupta dynasty were merely mahdrdjas, Chandragupta I (c. 319/320-c. 335 or 
c. 350), the son and successor of the second king, Ghatotkaca (c. 280-c. 319), assumed 
the title of mahdrdjadhirdja. This has led some historians to believe that the ancestors of 
Chandragupta I were petty landholders under the Later Kushans, the Bharashivas or the 

The Gupta era dates from the accession of Chandragupta I in c. 319/320, although the 
era itself was not introduced by him. Chandragupta I married a Licchavi princess early 
in his career. The Licchavis were an old-established clan who ruled over the Magadhan 
region during the first quarter of the fourth century. The Guptas were very proud of this 
alliance: they publicized it by issuing a class of gold coins known as the Chandragupta 
I-Kumaradevi type and by describing Samudragupta, the son and successor of Chan- 
dragupta I, as ' Licchavi-dauhitra' (son of the daughter of the Licchavis) in their inscrip- 

At the time of the death of Chandragupta I in c. 350, the Guptas, in alliance with the 
Licchavis, had become the greatest power of northern India. This alliance brought with it 
certain problems, however, since the nature and traditions of the two states were funda- 
mentally different. The Guptas were monarchical and patrons of Brahmanism, whereas the 
Licchavis had strong Buddhist leanings. The Allahabad pillar inscription tells us that Chan- 
dragupta nominated Samudragupta as his successor. This choice was obviously resented by 
some members of the family, since Kacha, who is known to us from his Chakradhvaja and 
Garudadhvaja variety of coins, revolted against his brother Samudragupta. Kacha's reign 
was shortlived, however; he was easily overcome and Samudragupta ascended the throne 
in c. 350. 

A lengthy eulogy to Samudragupta (who ruled until c. 375) was inscribed on an Asokan 
pillar at Allahabad that provides detailed information about his military achievements and 
lists the names of the states and people conquered by him. Unsupported by other evidence, 
and coming from a eulogy, this information must be treated with caution. Nevertheless 
the list is impressive. In real terms, however, Samudragupta's direct political control was 
confined to the Ganges valley, since the kings of the south and the Deccan were not under 
his suzerainty, but merely paid him tribute. The position was similar with the tribes of 
Rajasthan and Panjab, although Samudragupta's campaigns broke the power of the already 
weakened tribal republics. In the west, the Sakas remained unconquered. The validity of 
Samudragupta's wider claims is questionable. Daivaputra shdhi shdhdnushdhi is clearly 
a Kushan title, but the precise nature of the relationship with them remains uncertain (see 



ISBN 978-92-3-10321 1-0 Contents Origin and political history of the Guptas 

Chapters 5 and 6). Nevertheless Samudragupta achieved the difficult task of bringing about 
the political unification of the Ganges valley. 

Samudragupta was succeeded by his son, Chandragupta II, who ruled for 40 years 
(c. 375 - c. 415). There appears to have been trouble over his succession, just as in the case 
of his father. A play entitled the Devi Chandraguptam, written by Vishakhadatta some two 
centuries later and supposedly dealing with events on the death of Samudragupta, suggests 
that Ramagupta succeeded Samudragupta. The discovery of copper coins of Ramagupta in 
Vidisha-Airikina (in the eastern Malwa region), of the lion, garuda (a bird that was the vehi- 
cle of Vishnu and the badge of the Guptas), garudadhvaja (a garuda standard) and border 
legend types, lends credence to the possibility that Ramagupta was a governor of Malwa 
who assumed independence at the death of Samudragupta, but was eventually defeated by 
Chandragupta II. 

The Devi Chandraguptam, however, points to the fact that Chandragupta II's major 
campaign was fought against the Sakas. The Udaygiri cave inscription of Virasena, Chan- 
dragupta II's minister of war and peace, records that Chandragupta came with him to that 
region to 'conquer the whole world', referring to the Saka wars. The last known date of the 
ksatrapa coins is c. 388 and the earliest silver coins of Chandragupta II, struck in imitation 
of them, were of 409. Thus the annexation of western India to the Gupta kingdom must 
have taken place between these dates. This completed the Gupta conquest of northern India 
and gave them access to the western Indian ports. 

It is generally believed that Chandragupta II gave his daughter Prabhavatigupta in mar- 
riage to the Vakataka crown prince Rudrasena II to secure an ally for his Saka campaigns. 
But the Vakatakas, who had risen to the position of major power in the Vidarbha and adja- 
cent regions in the latter half of the third century, were then passing through a crisis and 
were thus unable to act as a safeguard for the Guptas against their Saka adversaries. The 
Guptas nevertheless put this marriage alliance to good use. Rudrasena II died five years 
after coming to the throne and as his sons were minors, his widow, the daughter of Chan- 
dragupta II, acted as regent from 390 to 410. This allowed the Guptas to secure virtual 
control of the Vidarbha region. 

Gupta power reached its apogee under Chandragupta II. In the east the frontiers were 
preserved and in the west they were stretched beyond the Jamuna. The republican states to 
the west of Mathura were finally integrated with the kingdom; western India was added; 
and the Deccan was brought under its orbit of direct influence. Chandragupta II assumed 
the title of Vikramaditya. He developed fully the concept of kingship, in consonance with 
the religious ideal of the time, as attested by the discovery of his Chakravikrama type of 
coins. The reverse of the coin contains a chakra (wheel), inside which is a standing male 



ISBN 978-92-3-10321 1-0 Contents Origin and political history of the Guptas 

handing three balls to a haloed royal figure. The entire symbol has been interpreted as the 
chakrapurusa of Vishnu, who is bestowing on the chakrabartl (sovereign) the three kingly 
virtues of authority, energy and counsel. 

The reign of Kumaragupta I (c. 415 - c. 454), the son and successor of Chandragupta II, 
was one of peace and relative inactivity. Thirteen inscriptions of his reign that have come to 
light show that, like his father, he succeeded in keeping the kingdom intact. The discovery 
of his coins from as far as Ahmedabad, Valabhi, Junagadh and Morvi suggests that he kept 
the newly acquired western provinces in a firm grip. There was possibly no fresh conquest 
to his credit. Towards the end of his reign, peace was disturbed by the invasion of an enemy 
whose identity has not been definitely established. According to the Bhitari pillar inscrip- 
tion of Skandagupta (c. 454 - c. 467), the son and successor of Kumaragupta I, the hostile 
forces belonged to a tribe called Pushyamitra. Far more serious, however, was the threat 
of a Huna (Hephthalite) invasion and Skandagupta had to concentrate on defending the 
kingdom against external invasions throughout his reign. Although the Bhitari inscription 
leaves no doubt as to the severity of the struggle, the Hunas were finally repulsed. 

After Skandagupta's death, the Guptas were unable to resist the repeated waves of Huna 
invasions (see Chapter 6) and central authority declined rapidly. The succession of the 
kings that followed is uncertain. A number of administrative seals have been discovered 
with the names of the same kings, but in a varied order of succession, which points to a 
confused close of the dynasty. A major blow came at the end of the fifth century, when the 
Hunas successfully broke through into northern India. 

The Hunas who attacked northern India, and eventually ruled parts of it, were not 
entirely independent but functioned under a Huna overlord whose dominions extended 
from Persia to Khotan. The Huna king Toramana consolidated Huna power in Panjab, 
from where he invaded the Gupta kingdom. Toramana was succeeded by Mihirakula, who 
ruled at the same time as the Gupta king, Narasimhagupta II, c. 495. In his struggle against 
Mihirakula, Narasimhagupta II received support from some powerful feudatories, partic- 
ularly the Maukhari chief Ishvaravarman and Yashodharman of Malwa, whose Mandasor 
inscription states that Mihirakula paid tribute to him. The political impact of the Hunas in 
India subsequently subsided. Acting as a catalyst in the political process of northern India, 
however, the Hunas saw the slow erosion and final dissolution of the Gupta kingdom by 
the middle of the sixth century. 

With the disintegration of the Gupta kingdom, the notion of a pan-Indian Empire came 
to an end until the advent of the Turks, although it was briefly revived during the reign of 
Harshavardhana in the seventh century. The post- Gupta period in northern India saw the 
emergence of regional kingdoms, mostly derived from the feudatories of the Guptas. The 



ISBN 978-92-3-10321 1-0 Contents Origin and political history of the Guptas 

more important among them were the Later Guptas, the Maukharis, the Pushyabhutis and 
the Maitrakas. 

The Later Guptas had no connection with the Gupta main line. The Aphsad inscrip- 
tion gives a detailed history of the dynasty which shows that the Later Guptas were rulers 
of Magadha with suzerainty over Malwa. They were eventually ousted from Magadha by 
the Maukharis of Kanauj, who originally held the region of western Uttar Pradesh. The 
Pushyabhutis ruled in Thane swar (modern Harvana). They had made a marriage alliance 
with the Maukharis and on the death of the last Maukhari king, the Maukhari nobles 
requested Harsha, the reigning king of the Pushyabhuti dynasty, to unite his kingdom with 
them and rule from Kanauj. The Maitrakas ruled in Gujarat, with Valabhi as their capital. 

Of all these states which arose out of the ruins of the Gupta kingdom, that of Valabhi 
proved to be the most durable. The unusually large number of records of this family that 
have come to light help to reconstruct their political history with some degree of certainty. 
There were able rulers among them, such as Shiladitya, under whose leadership Valabhi 
became the most powerful kingdom of western India towards the close of the sixth cen- 
tury. The Maitrakas continued to rule until the middle of the eighth century, when they 
succumbed to outside attacks - probably from the Arabs, as mentioned by al-Biruni. 

Of all the successor states to the Guptas, that which rose to greatest eminence, however, 
was ruled by the Pushyabhutis of Thaneswar. The Pushyabhuti family came to the fore with 
the accession of Prabhakaravardhana, but it was during the reign of his son Harshavard- 
hana (606-647) that they succeeded in establishing political authority over most parts of 
northern India. The early history of Harsha's reign is reconstructed from his biography, the 
Harm Caritam, written by his court poet Bana. This is supplemented by the account of 
the Chinese pilgrim Hsiian-tsang, who visited India during Harsha's reign. Harsha made 
Kanauj the seat of his power and it rose to political prominence from the late sixth century 
as a place of strategic importance. From there he extended his authority in all directions. 
Rajasthan, Panjab, Uttar Pradesh, Bihar and Orissa were all under his direct control and 
he exercised influence over a much wider area. The peripheral states acknowledged his 
suzerainty and thus Harsha, like the Guptas, ruled a large kingdom in northern India that 
was loosely connected by feudal ties. 

The most important political development in western India from the seventh century 
was the rise of the Rajputs. Their origin is somewhat obscure, but it has been suggested 
that they came from Central Asia with the Hunas, displaced the original tribal inhabitants 
of Rajasthan and laid the foundation of the later Rajput families. The theory of indigenous 
origin has also been proposed. The most notable among the Rajput dynasties were the 



ISBN 978-92-3-10321 1-0 Contents Social and economic conditions 

Gurjaras Pratiharas, the Guhilas and the Cahamanas, but they were to play their part in 
wider Indian politics only at a later date. 

Social and economic conditions 

For a reconstruction of social conditions under the Guptas, we depend heavily on the con- 
temporary legal texts, or smrtis. A number of such texts, most of which took the Dhar- 
masastra of Manu as their basis, were written during this period, the best-known being 
the Yajnavalkya, the Ndrada, the Brhaspati and the Katyayana. These smrtis provide an 
ideal representation of society from the brahmanical point of view. Contemporary Sanskrit 
plays and prose literature, however, do not always corroborate this ideal and it may be 
safely assumed that the injunctions of the smrtis were not necessarily strictly enforced. 
This conclusion is supported by the inscriptions of the period and by the accounts of the 
Chinese pilgrims Fa-hsien and Hsiian-tsang. 

In the Gupta period, brahmanical reaction against Buddhism and Jainism became 
stronger. As a result, varna- (i.e. caste-) based social stratification and the supremacy of 
the brahmans (the highest caste) received much greater emphasis. It is difficult to ascertain 
the caste of the Guptas, but they were, in all probability, brahmans themselves and strongly 
supported the brahmanical social order. The brahmans were given land on a large scale 
and they claimed many privileges which are listed in the Ndrada. For example, under no 
circumstances was capital punishment to be inflicted on them or their property confiscated. 
The ksatriyas (the second, or warrior, caste) continued to enjoy great prestige due to their 
political influence, and there was a tacit understanding between these two upper castes in 
sharing social and political power. 

The degeneration of the vaisyas (the third, or trader, caste), which had begun earlier, 
intensified during this period. Because of advanced agricultural techniques and 
developments in handicrafts, the condition of the sildras (the fourth, or menial, caste) 
improved and there was no great difference between a poor vaisya and a prosperous sudra. 
The vaisyas, however, retained their supremacy in industry and commerce and held impor- 
tant positions on the municipal boards. There are repeated references to the sudra peasantry 
in the contemporary sources as opposed to their status as agricultural labourers in earlier 
times. The smrtis of the Gupta period make a clear distinction between the sudras and the 
slaves. This period saw the emergence of the untouchables, who were beyond the pale of 
the caste structure and lived outside the city boundaries. 

From this cumulative evidence it appears that the significance of the traditional varna 
structure, based on colour and race, was being seriously undermined and the jati 



ISBN 978-92-3-10321 1-0 Contents Social and economic conditions 

structure, based on occupational status, was becoming increasingly important. Like the 
varnas, the jati system was hereditary and the number ofjatis gradually proliferated. As a 
social institution the jatis were independent of the varnas, although Hsiian-tsang describes 
occupations demarcated for each of the four varnas. In this period the jati system was 
not particularly strict and it was still possible for a person to move from one occupational 
status to another. That social mobility was not altogether restricted is demonstrated by 
examples of brahmans taking up the professions of merchant, architect or government offi- 
cial. Hsiian-tsang gives a comparative account of the political rights of the four varnas. 
He had seen five brahman, five ksatriya, two vaisya and two sudra kings. However, people 
increasingly came to be identified with the small occupational groups and the wider varna 
consciousness was replaced by a commitment to the jatis. 

The brahmans had tried to explain the creation of the jatis in terms of the mixed castes, 
born out of intermarriage between the varnas,, which was prohibited but practised. The 
father of Bana married a sudra woman. The Yajnavalkya prescribed that the son of a sudra 
mother and a brahman father should inherit his father's property, although this right was 
not recognized in the Brhaspati, a text composed towards the end of the Gupta period. The 
contemporary smrtis mention a number of mixed castes. 

Although women were idealized in literature and art, in practice they had a distinctly 
subordinate social position. Education of a limited kind was permitted to upper-class 
women but they were not allowed to participate in public life. Early marriage was advo- 
cated and strict celibacy was recommended for widows. The attitude of the contemporary 
smrtis towards women was one of contempt. Women were described as almost a consumer 
commodity, exclusively owned by their husbands. But there were exceptions to this norm 
in real life. For example, as mentioned earlier, Prabhavatigupta, the daughter of Chan- 
dragupta II, managed the affairs of state for some 20 years. On the whole, however, the 
only women to enjoy a measure of freedom were those who deliberately chose to opt out 
of the prevailing system of regulations by becoming a Buddhist nun or a courtesan. 

The social supremacy of the brahmans is also reflected in the economy of the period, 
as attested by the frequency of tax-free land-grants made to them. This was a period of 
partial decline in trade and consequently a greater concentration on land. There were four 
categories of land - fallow and waste land, state-owned land and privately owned land. 
Agriculture expanded with the reclamation of new land for cultivation. Contemporary texts 
reveal a more liberal and practical attitude towards waste land, with the state encouraging 
the peasantry to bring uncultivated and forest land under the plough. Those who reclaimed 
land on their own initiative and made arrangements for its irrigation were exempted from 
paying taxes until they started earning an income of twice their original investment. Inscrip- 



ISBN 978-92-3-10321 1-0 Contents Social and economic conditions 

tions of the Gupta period repeatedly mention the sale and purchase of waste land, which 
indicates that such transactions were financially profitable. The state actively patronized 
agricultural activity. This is suggested by the Junagadh inscription of Skandagupta, which 
records work on Lake Sudarsana at Girnar under state supervision, presumably for irriga- 
tional purposes. Kalidasa describes agriculture and animal husbandry as the mainstay of 
the royal exchequer, since the major portion of revenue came from the land, at one-sixth of 
the net produce. 

Agricultural implements remained much the same, although iron was more widely used 
for their manufacture. Varhamihira, in his astrological work, the Brhat-samhitd, refers to 
an instrument for measuring rainfall. Crops were grown twice a year. According to Hsiian- 
tsang, sugar cane and wheat were grown in the north-west and rice in Magadha and further 
east. Southern India was known for black pepper and spices. The Amarakosa, the San- 
skrit lexicon belonging to this period, also refers to a large variety of fruit and vegetables. 
Despite overall growth, however, brahmanical and Buddhist religious injunctions were not 
conducive to the expansion of agriculture. The Brhaspati was unwilling to respect the 
income derived from agriculture and cultivation was prohibited for the Buddhist monks. 

The manufacture of textiles of various kinds was one of the more important industries 
at this time. There was a vast domestic market, since textiles were a prime item of trade 
between northern and southern India. There was also a considerable demand in foreign 
markets. Silk, muslin, calico, linen, wool and cotton were produced in great quantity. The 
production of silk decreased towards the end of the Gupta period since many members of an 
important guild of silver-weavers in western India abandoned their traditional occupation 
and took to other professions. This might have been due to the increasing use of the Silk 
Route and the Sea Route to China, which brought a large amount of Chinese silk to India 
or, more generally, to the decline in trade with the West. Metalwork, particularly in copper, 
iron and lead, continued as one of the essential industries. The use of bronze increased and 
gold and silver ornaments were in constant demand. We have little clue as to the sources 
of the abundant supply of metals in the Gupta period and it seems that copper, lead and tin 
had to be imported from abroad. Gold may have been obtained from the Byzantine Empire 
in exchange for Indian products, although Hsiian-tsang mentions that it was also produced 
indigenously in huge quantities. The working of precious stones continued to maintain its 
high standard. Pottery remained a basic part of industrial production, although the elegant 
black polished ware of earlier times was now replaced by an ordinary red ware with a 
brownish slip. 

The guild was the major institution in the manufacture of goods and in commercial 
enterprise. Some historians believe that the importance of the guilds declined in the Gupta 



ISBN 978-92-3-10321 1-0 Contents Social and economic conditions 

period. India no longer participated in the long-distance trade in luxury goods. Instead a 
new kind of commercial network emerged on regional lines, based on the exchange of arti- 
cles in daily use. In these changed circumstances, the powerful guilds of the earlier times 
disintegrated. Contemporary sources, particularly the seals found at Vaisali and Bhita, sug- 
gest nevertheless that both the activities and the significance of the guild remained during 
this period. Guilds sometimes acted as bankers and loaned money on interest, as did some 
of the Buddhist sanghas (communities). The rate of interest varied according to the pur- 
pose for which money was required. The lowering of the interest rate implies an increased 
confidence in overseas trade as well as a greater availability of goods and the consequent 
decrease in profit margins. 

Trade between northern India and South-East Asia was conducted through the ports 
of the east coast. The west coast ports served as the link in India's trade contacts with 
the Mediterranean region and Western Asia. Several inland routes connected India with 
China through Central Asia and Tokharistan and across the Karakorum range and Kashmir. 
The most important event in the economic history of East and South-East Asia during 
this period was the development of an inter-oceanic trade, reaching from China through 
Indonesia and the east coast of India up to Simhala and extending from there along the 
west Indian coast to Persia, Arabia and Ethiopia. Despite commercial competition between 
China and India, the two countries maintained close links. Coins of the T'ang emperors of 
China have been discovered in southern India and Indian merchants resided in Canton. Still 
more far-reaching in their consequences were India's trade contacts with South-East Asia, 
leading to Indian settlements there and an Indian influence that permeated the local pattern 
of life, particularly in Thailand, Cambodia and Java. 

The export of spices, pepper, sandalwood, pearls, precious stones, perfumes, indigo and 
herbs continued as before. Pepper was exported from the ports of the Malabar coast and 
sesame, copper and cotton garments from Kalyana. The Pandya area had an important role 
to play in the pearl trade. The commodities that were now being imported to India, however, 
differed from those in earlier times. Chinese silk came in greater quantity, as did ivory 
from Ethiopia. Imports of horses from Arabia, Iran and Tokharistan also increased. Copper 
came from the western Mediterranean region and sapphire from Simhala. The Gupta king 
issued special charters to merchants' organizations which relieved them of government 
interference. Since this was the time when the law-makers declared it a great sin for a 
brahman to travel by sea, this may have resulted in reduced Indian participation in maritime 

Some historians have characterized the socio-economic developments of the Gupta 
period in terms of feudalism. They argue that although there had been a long tradition of 



ISBN 978-92-3-103211-0 Contents Administration 

donating land to the brahmans, the number of such donations greatly increased in the Gupta 
period. Villages along with their inhabitants, revenue due to the king, administrative and 
judicial rights, exemption from the interference of government officials, and even the right 
to enjoy fines levied on cultivators, were all transferred to the religious beneficiaries. What 
began as grants to the priestly class were later extended to administrative officials. With the 
emergence of a local, self-sufficient economy, religious donations as well as land-grants to 
secular officials (either in lieu of salary or as a reward for services) became popular. The 
principal characteristics of this selfsufficient economy were the decline of trade and urban 
centres and a scarcity of coinage. Thus from the economic point of view, the central feature 
of Indian feudalism was the emergence of landed intermediaries. As a result, the freedom 
of the peasantry was curtailed, their mobility was restricted and they were forced to serve 
as unpaid labour. 

Those historians who do not subscribe to this view have challenged the premises of 
Indian feudalism. They argue that during the Gupta period, trade did not decline and the 
scarcity of coins was at best marginal. Quantitative analyses of the coinage of this period 
have still to be made and the relative scarcity of coins is still merely an assumption. Some 
of the old-established towns did lose their importance, but new urban centres emerged to 
replace them. Finally, the two indispensable institutions of European feudalism, namely 
manor and serfdom, never developed in India. Historians who subscribe to this second 
view are therefore inclined to describe the practice of land-grants as nothing but India's 
traditional landlordism. The debate is still to be settled. 

The literary records of this period suggest an overall economic prosperity at least among 
the upper classes. Fa-hsien describes the people of Madhyadesha (the 'middle country') as 
prosperous and happy towards the beginning of the fifth century. Evidence of material con- 
ditions obtained from excavations also points to a high standard of living. The prosperous 
urbandwellers lived in luxury; and comfort, in the urban centres at least, was not confined 
to the upper classes. Yet it was a culture with wide variations. The untouchables lived on 
the outskirts of the opulent cities and the peasantry were being gradually impoverished. 
The maintenance of an imperial facade was a purposeless expense which must have been 
a drain on the economy. Indeed, the debased Later Gupta coinage indicates an economic 


In many respects, the Gupta administration constitutes the watershed between India's past 
and future traditions of polity and government. The most noticeable feature of the 



ISBN 978-92-3-103211-0 Contents Administration 

post-Mauryan administrative development was the gradual erosion of the government's 
centralized power. First, the Satavahanas and the Kushans entered into feudatory rela- 
tions with the smaller kingdoms. Second, land-grants, which began from this time, created 
administrative pockets in the countryside managed by the religious beneficiaries. A third 
factor which contributed to the process of decentralization was the existence of autonomous 
governments in several cities of northern India. Guilds of traders from these cities even 
issued coins, which was normally the prerogative of the sovereign power. At several points, 
however, the old centralized system of administration was continued and even strengthened 
by the accession of new elements. 

The Guptas discarded the modest title of raja and adopted the high-sounding ones 
brought into vogue by the Kushans. The most typical example is mahdrdjadhirdja which, 
along with its several variants, appears in Gupta inscriptions. The Gupta kings also claimed 
superhuman qualities for themselves. They continued the traditional machinery of bureau- 
cratic administration with nomenclature that was mostly borrowed or adopted from ear- 
lier times. Thus the mantri (prime minister) stood at the head of the civil administration. 
Among other high officers were the mahdbalddhikrta (commander-in-chief), 
mahddandandyaka (general) and mahdpratihdra (chief of the palace guards). A high- 
ranking officer, encountered for the first time in the Gupta records but destined to have 
a long career, was the sandhivigrahika (foreign minister). The bhuktis (provinces) were 
usually governed by princes of royal blood and sometimes by a class of officers called 
uparikas. The link between the central and provincial administration was furnished by 
kumdrdmdtyas and dyuktas who ruled over visayas (districts). The district officers were 
nominated by the provincial governors. 

For the first time, the inscriptions give us an idea of systematic local administration in 
the Gupta period, which assumed many new dimensions. The series of northern Bengal 
epigraphs mentions the adhisthdnddhikarana (municipal board), visayddhikarana (district 
office) and astakulddhikarana (possibly, rural board). The full adhisthdnddhikarana is said 
to consist of four members, the nagarasresthl (guild president), the sdrthavdha (chief mer- 
chant), the prathamakulika (chief artisan) and the prathamakdyastha (chief scribe). The 
precise significance of the astakulddhikarana is unknown, but in one example it is said 
to be headed by the mahdttaras (village elders) and also includes the grdmika (village 
headman) and the kutumbins, (householders). 

Under the Guptas, the scope and functions of royal authority underwent a significant 
change. The Guptas left a number of conquered states in a position of subordinate inde- 
pendence. With the exception of Uttar Pradesh, Bihar and parts of Bengal, the kingdom 
was held by feudatories such as the Parivrajaka princes, who issued their own land-grants. 



ISBN 978-92-3-103211-0 Contents Religious life 

The presence of these feudatories must have severely restricted the Guptas' royal author- 
ity. We do not have much information about military affairs, but can reasonably surmise 
that the troops supplied by the feudatories must have accounted for a good proportion of 
the Gupta army. The state no longer enjoyed a monopoly over the possession of horses and 
elephants. The significant aspect of Gupta bureaucracy was that, since it was less organized 
and elaborate than the Mauryan administration of the third century B.C. (seen in Kautilya's 
Arthasastra), it allowed several offices to be combined in the hands of the same person and 
posts tended to become hereditary. In the absence of close supervision by the state, vil- 
lage affairs were now managed by leading local elements who conducted land transactions 
without consulting the government. 

Similarly in urban administration, organized professional bodies enjoyed considerable 
autonomy. The law-codes of the Gupta period, which provide detailed information about 
the functioning of the guilds, even entrusted these corporate bodies with an important share 
in the administration of justice. With the innumerable jatis (which were systematized and 
legalized during this period) governing a large part of the activities of their members, very 
little was left for central government. Finally, the Gupta kings had to take account of the 
brahman donees, who enjoyed absolute administrative privileges over the inhabitants of 
the donated villages. Thus in spite of the strength of the Gupta kings, institutional factors 
working for decentralization were far stronger during this period. This Gupta adminis- 
tration provided the model for the basic administrative structure, both in theory and in 
practice, throughout the early medieval period. 

Religious life 

The rise of the Guptas was analogous to the emergence of Puranic Hinduism. The vehi- 
cle for the propagation of this resurgent Hinduism was a set of texts called the Purdnas, 
the earliest of which were composed in this period. The Purdnas, which began as the 
historical tradition recording the creation of the universe and detailed the genealogies of 
each dynasty, were originally composed by bards. During this period, however, they were 
rewritten by the brahmans in classical Sanskrit to include information on Hindu sects, rites 
and customs. Before the coming of the Guptas, the ideal brahmanical social order had 
been disrupted to such an extent by rulers who patronized the heretical cults that we see an 
obsessive fear of the Kali, or Dark Age, in all the early Purdnas,. 

All the major aspects of brahmanical religion, by which Puranic Hinduism came to be 
identified in later centuries, crystallized in this period. The image of the deity emerged as 
the centre of worship and worship superceded sacrifice, although a sacrificial offering to 



ISBN 978-92-3-103211-0 Contents Religious life 

the image remained central to the ritual. This in turn encouraged bhakti (devotionalism), 
which consisted of an intense personal attachment to the object of worship. As a result, 
worship of a god became an individual concern and the priest ceased to be so dominant a 
figure as in the sacrifice. 

Hindus became divided into two main sects, Vaishnava and Shaiva, claiming Vishnu 
and Shiva respectively as the supreme deity, just as each Pur ana extolled the superiority 
of one or the other. The worshippers of Vishnu were more prevalent in northern India, 
where they received active patronage from the Guptas; Chandragupta II called himself a 
paramabhagavata (devotee of Vishnu). Shaivism took firm root in the south, although it 
was not confined to that region. The Huna king Mihirakula, Shashanka the ruler of Bengal, 
some kings of the Pushyabhutis of Kanauj and the Maitrakas of Valabhi were all followers 
of Shiva. Despite such sectarian preferences, at times expressed in acute rivalry, there was 
an underlying strain of monotheism in Puranic Hinduism which saw the various deities 
as manifestations of a unified whole. The social existence of a Hindu came to be defined 
in terms of a correct dharma (law), artha (economic well-being), kama (sensual pleasure) 
and moksa (salvation of the soul). 

A notable feature of intellectual life in this period was provided by the lively philosoph- 
ical disputations between the Buddhists and the brahmans, centring around six different 
schools of thought which came to be called the six systems of Hindu philosophy. Although 
their origin can be traced to the thinking of a much earlier period, some of their cardinal 
principles were enunciated at this time. Vedanta is the most influential of the six systems. 
The doctrines of Vedanta were based on the Upanisadas (books of the teaching of sages) 
and gave logical and organized form to their many mystical speculations. It postulated the 
existence of the Absolute Soul' and maintained that the final purpose of existence was the 
union of the individual and this Absolute Soul' after physical death. Together these six 
systems constitute the core of Hindu philosophy and all subsequent developments are its 

Although Buddhism was theoretically still a formidable rival of Hinduism, by the end 
of this period its influence was waning (see Chapter 18, Part Two). 

Jainism was saved from a similar fate by its essentially conservative character. Unlike 
the other religious systems, it underwent little change in ideas or doctrines. The fact that it 
failed to adapt to new environments accounts for its restricted popularity but much longer 
life compared with Buddhism. Jainism continued to be supported by the merchant com- 
munity of western India. In certain areas of the Deccan and the south it received patron- 
age from local royalty, though much of this patronage ceased after the seventh century. 
The organizational split between the two principal Jaina sects, the Svetambaras and the 



ISBN 978-92-3-103211-0 Contents Literature 

Digambaras, reached its culmination during this period. In the early sixth century, the 
second Jaina Council was held at Valabhi to recover and systematize the Jaina canonical 
instructions which were facing extinction. At this council, the Jaina canon was defined 
substantially as it exists today. The Jainas had by now evolved a series of icons: the images 
of the tirthankaras (Jaina teachers) in the Khandagiri cave at Bhubaneshwar are some of 
the best examples. 


Sanskrit literature was given lavish encouragement during this period, mostly through 
royal patronage. It was a literature of the elite and those associated with the court cir- 
cle. Classical Sanskrit poetry flourished with Kalidasa's works probably in the late fourth 
and early fifth centuries. Kalidasa reflects the court culture of the time. Though deeply 
imbued with tradition, all his works reveal his distinct personality. He wrote two long 
poems, the Kumarasambhava and the Raghuvamsa, and also the Meghaduta, a work of a 
little over 100 verses, which is one of the most popular Sanskrit poems; it has unity, balance 
and a sense of wholeness that is rare in early Indian literature. Kalidasa's long poem the 
Kumarasambhava has a religious theme, but is essentially secular in character and contains 
passages of great beauty. 

Many poets after Kalidasa wrote courtly epics, but none so ably as he. The two best 
examples of such poems are Bharavi's Kiratarjumya (mid-sixth century) and Magha's 
Sisupalavadha (late seventh century). Magha had set the trend for the poetic style of the 
later period, which became progressively ornate and artificial. The finest poet in this genre 
was Bhartrhari, possibly of the seventh century, who left only 300 separate stanzas on the 
subjects of wordly wisdom, love and renunciation respectively, which are considered mas- 
terpieces of concise expression. Another important exponent of this style was Amaru, also 
of the seventh century. 

As in poetry, the greatest exponent of Sanskrit drama in this period was Kalidasa, who 
was able to achieve the effects he wanted and to capture the conflicting emotions of his 
characters. The real value of his work, however, lies in his imagery, language and dialogue, 
which are fresh and vigorous. Shudraka, probably Kalidasa's contemporary, has left only 
one play, the Mrcchaktika, which is the most realistic of Indian dramas. Vishakhadatta, 
who probably belonged to the sixth century, has only one complete surviving play, the 
Mudraraksasa; the plot is exceedingly complicated, but is worked out with great skill and 
leads to a breathtaking climax. One interesting convention of the Sanskrit theatre of this 
period is that it allows no tragedy. Tragic and pathetic scenes are common enough but the 



ISBN 978-92-3-103211-0 Contents Science 

endings are almost invariably happy and melodramatic, often necessitating an unnatural 
forcing of plots. Another notable feature is that the characters of high social status speak 
Sanskrit while women and the 'lower orders' speak Prakrit: this defines the standing of 
Sanskrit and Prakrit in a social context. The best examples of Sanskrit prose literature of 
this period are provided by Dandin, Subandhu and Bana, all of whom lived in the late sixth 
and early seventh centuries. 

The Svetdmbara Jaina canon and its exegetic literature in Ardha-Magadhi Prakrit, the 
few religious texts of the Digambara Jainas in Shauraseni Prakrit and the commentaries of 
Buddhist texts written in Pali constitute the most important specimens of Prakrit and Pali 
literature of this period. The attempts of the Jaina monks to redefine their canon, following 
the second Jaina Council, resulted in the production of a vast literature, which is didactic 
in style, arid in content and deficient in literary value. Mention may also be made of inde- 
pendent religious narratives such as the Vasudevahindi by Dharmadasa and Sanghadasa 
and a religious romance called the Tarangavartlkathd attributed to Padalipta. Among the 
Prakrit long narrative poems, the most noteworthy are the Setubandha by Pravarasena and 
the Gauda-vaho by Vakpatiraja. 

The Gupta period is referred to as the ' classical age' of ancient India, mainly because of 
its cultural achievements. The description seems to be true for the upper classes, amongst 
whom material and intellectual culture reached a level never before attained. It has been 
suggested that every great literary form implies the unfolding of a new social grouping, 
headed by some new class. Those who hold this view argue that this great period of clas- 
sical Sanskrit literature - which witnessed an unprecedented growth and development - 
was intimately connected with the rise of feudalism. Motivated by an entirely different set 
of reasons, the nationalist historians of the early twentieth century sought instead to locate 
the Utopian ' golden age' in this period, again primarily because of its literary and artis- 
tic excellence. These divergent conclusions, however, agree on the common point of the 
cultural flowering during this period. 


There was a corresponding development in the field of science, though it was not com- 
parable in scale or quality with the growth in literature, and the knowledge of metals had 
improved tremendously. The treatises of Astdnga-samgraha and Astdnga-hrdaya-samhitd 
were mostly compilations from earlier texts. Books on the diseases of animals, particularly 
horses and elephants, now appeared for the first time. 



ISBN 978-92-3-103211-0 Contents Art and architecture 

It was an intensely active period in mathematics which encouraged the development of 
astronomy as a precise science. Aryabhata, who composed his famous work the Aryabhatiya 
in 499, was an accomplished mathematician who knew the use of the decimal place-value 
system and dealt with area, volume, progressions, algebraic identities and indeterminate 
equations of the first degree. He was the first writer to hold that the earth was a sphere 
rotating on its axis and that eclipses were caused by the earth's shadow falling on the 
moon. With remarkable accuracy, Aryabhata calculated the length of the solar year to be 
365.3586805 days. Varhamihira, who is more known for his astrological work the Brhat- 
samhita, flourished in the sixth century. 

Despite an accurate knowledge of the duration of the solar year, the basic unit in record- 
ing dates was the lunar day, approximately 30 of which formed the lunar month. Twelve 
lunar months make only 354 days and hence every 30 months an extra month was added to 
the year. The Hindu calendar, though quite accurate, was thus rather cumbrous. The solar 
calendar, imported with Western astronomy, was also known from the Gupta period, but it 
did not replace the lunar calendar. Hindu thinkers had evolved a cyclical theory of time. 

Art and architecture 

The Gupta period also represents a watershed in the history of Indian art. In one respect, it 
marks the culmination and ultimate exhaustion of earlier tendencies in architectural types 
and forms. In another, it marks the beginning of a new age, connected with the phenomenal 
growth and development of the temple. The material prosperity of the period is reflected in 
its town planning. Most cities were laid out in squares; wooden buildings were replaced by 
buildings of brick; houses were oriented to the cardinal points; and drains and wells were 
carefully planned. 

Rock-cut cave architecture persisted in this period - mostly Buddhist but with a few 
brahmanical and Jaina examples. The rock-cut architecture of the Buddhists consisted of 
two conventional types, the chaitya (shrine containing a stupa) and the vihdra (monastery). 
The most notable groups are found at Ajanta, Ellora, Aurangabad and Bagh. Of the 23 
caves at Ajanta which were excavated during this period, only caves XIX and XXVI were 
chaitya halls and the rest were viharas. The most significant innovation here is the wealth 
of sculptures of the human figure. 

The vihdra was planned in the form of rows of cells around a central court. Of the 
viharas at Ajanta, the most important are caves XVI, XVII, I and II, remarkable for the 
beauty of their pillars. Of the stupas, which were built in large numbers, two deserve special 
mention - that at Mirpur Khas and the Dhamekh stupa at Sarnath. The rich, elegant patterns 



ISBN 978-92-3-103211-0 


Art and architecture 

of the ornamental scheme constitute the chief beauty of the Sarnath monument and its 
cylindrical shape indicates a date of c. the sixth century. 

Unfortunately not much has survived of Gupta temple architecture, although the sources 
indicate that many temples were constructed. It has been suggested that such temples were 
on the whole unimpressive shrines which were either absorbed in domestic architecture or 
else built over in later centuries. Extant examples consist of three major groups - the flat- 
roofed square temple with a shallow porch in front or a squat tower above; the rectangular 
temple with an apsidal back and barrel-vaulted roof; and the circular temple with shallow 
projections at the four cardinal points. The Dashavatara temple (fifth century) at Deogarh 
(Figs. 1 and 2) is one of the best examples of an age of experiments in types and forms 
which was later elaborated and finally crystallized in the eighth-century Hindu temple in 
northern India. 

The pivot of Gupta sculptural art was the human figure. By now all animal and vegetal 
patterns had been eliminated from the narrative and simply underlined the importance of 
the human form. The body was given perceptual form with the help of a full modelling 
that, in its naturalism, is almost unparalleled in Indian art. The Buddhas and Bodhisattvas 

FIG. 1. Deogarh. Dashavatara temple. General view. (Photo: © Archaeological Survey of India, Jan- 
patch, New Delhi.) 



ISBN 978-92-3-103211-0 


Art and architecture 

FIG. 2. Deogarh. Dashavatara temple. Side view. (Photo: © Archaeological Survey of India, Janpatch, 
New Delhi.) 

of the fifth and sixth centuries represent the final achievement of the highly subtle, mys- 
tical and fluid thought of the Mahayana school. The most important centres of sculpture 
were Mathura and Sarnath. One of the best examples is the seated Buddha in dharma- 
chakrapravarttana (Fig. 3) attitude from Sarnath, where the body sheds its toughness and 
attains complete ease and serenity. All this is achieved with the help of delicate modelling, 
a smoothly flowing, melting line and an utmost economy of technique. 

The Hindus, however, treated the image as a symbol. Although the god took a human 
form, he might well have several arms or the head of an animal. The Hindu gods, as repre- 
sented in the sculpture of this period, were mainly incarnations of Vishnu, the most popular 
among them being those of nrsimha (half man/half lion) and vardha (boar). The cult of 
Shiva was mostly confined to phallic worship, which did not offer much sculptural scope. 
The more significant brahmanical sculptures of the time were influenced by the Puranic 
vision of the evolution of the universe from its material cause and its re-creation from the 
constituent elements into which it is merged. This explains the origin and meaning of the 
latent dynamic strength and power in the magnificent reliefs of the Udaygiri caves of east- 
ern Malwa or of Badami, Ellora, Aurangabad and Elephanta. For example, the Great Boar 



ISBN 978-92-3-103211-0 


Art and architecture 

FIG. 3. Sarnath. Buddha in dharma-chakra-pravarttana. (Photo: © Archaeological Survey of India, 
lanpatch, New Delhi.) 

(an incarnation of Vishnu who rescued the earth from the cosmic ocean) carved in relief 
near the entrance of a cave at Udaygiri conveys the impression of a great primordial power 
working for good against the forces of chaos and destruction, and bears a message of hope, 
strength and assurance. 

While the quest for form in stone concerned itself with themes and expressions of 
a deeper and more fundamental significance, painting had a secular character and was 
presumably in more general demand. The Visnudharmottara, a text of the Gupta period, 



ISBN 978-92-3-103211-0 


Art and architecture 

devotes an entire chapter to the art of painting, laying down many of its theoretical canons. 
The best examples of painting can be found in the murals of caves I, II, XVI, XVII and 
XIX of Ajanta (Figs. 4-7), caves IV and III of Bagh and caves III and II of Badami. The 
Ajanta paintings do not show a progressively developing style, as in contemporary sculp- 
ture. The murals chiefly depict scenes from the life of the Buddha and from the jatakas 
(birth stories of the Buddha). There is no perspective, but an illusion of depth is given by 
placing the background figures somewhat above those in the foreground. Although painted 
for religious purposes, the Ajanta murals bear a secular message. They depict the entire 
panoply of life in ancient India: princes in their palaces, ladies in their apartments, coolies, 
beggars, peasants and ascetics, together with the many Indian birds, beasts and flowers. 

Very different are the enormous number of terracotta reliefs from northern India and 
Bengal. Produced from sketchy moulds in large quantities, they were carefully finished 
and often painted. Employed for various purposes, their primary use was in decorating the 
exterior walls of Buddhist establishments and residential houses. 

A remarkable example of handicraft is the ivory Triratna (trident symbol of the three 
jewels of Buddhism) in high relief representing a Buddha with attendant Bodhisattvas. The 
central figure is like a miniature Mathura image of the period. The wealth of jewellery worn 
by women of this period is seen in the flying apsaras (nymphs of the sky) in the Ajanta 

FIG. 4. Ajanta. Mural painting in cave XVI. Dying princess. (Photo: © Archaeological Survey of India, 
Janpatch, New Delhi.) 



ISBN 978-92-3-103211-0 


Art and architecture 

FIG. 5. Ajanta. Mural painting in cave XVII. Indra accompanied by his celestial musicians. (Photo: 
© Archaeological Survey of India, Janpatch, New Delhi.) 

FIG. 6. Ajanta. Mural painting in cave XVII. Vessantara y'ata/M. (Photo: © Archaeological Survey of 
India, Janpatch, New Delhi.) 



ISBN 978-92-3-103211-0 


Art and architecture 

FIG. 7. Ajanta. Mural painting in cave XVII. Rahul and his mother Yashodhara. (Photo: © Archaeo- 
logical Survey of India, Janpatch, New Delhi.) 

- IBVsi > 


FIG. 8. Ajanta. Mural painting in cave XVI. Flying apsara. (Photo: © Archaeological Survey of India, 
Janpatch, New Delhi.) 

murals (Fig. 8), which also show the variety of high-quality textiles such as embroidery, 
tie and dye work, brocade and muslin. A rare example of Gupta metalwork is an object 
that has been identified as an architect's plummet, made of iron coated with bronze. On its 



ISBN 978-92-3-103211-0 Contents Art and architecture 

neck is a plaque with a representation of dancing figures framed in prongs terminating in 
lotus buds which is reminiscent of the decorative forms of Gupta stone sculpture. Among 
the most splendid examples of the minor arts of the period are the gold coins of the Gupta 
dynasty, executed with impeccable finesse. 

Some authorities have depicted the Gupta kings as the liberators of India from foreign 
rule. But the invaders had become thoroughly Indianized by the Gupta period and this 
made the task of assimilating them into Indian society relatively simple by assigning them 
appropriate caste status. They continued to exert an influence on aspects of Gupta culture, 
however, and this is nowhere more pronounced than in Gupta art. Many characteristic 
architectural forms and motifs of the Guptas were inherited from Kushan Mathura and 
Gandhara. Gupta sculpture undeniably developed from an emphasis on massive power 
inherited from the Kushans, but gradually it evolved its own style, with graceful and more 
linear creations. Gupta culture, with all its inevitable borrowings from previous traditions, 
was essentially indigenous in character and set the norms for subsequent developments. 



ISBN 978-92-3-103211-0 Contents KHWARIZM 


E. E. Nerazik and P. G. Bulgakov 



Uncertain early history 212 

Social structure and administration 215 

Art, architecture, religion and language 222 


The Khwarizmian calendar 227 

Khwarizmian eras 229 

Religious beliefs 230 

The Arab conquest 233 

See Map 5. 



ISBN 978-92-3-103211-0 Contents Uncertain early history 

Part One 


(E. E. Nerazik) 

Uncertain early history 

The period from the late third to the eighth century was a very complex one in the history of 
Khwarizm (Chorasmia). It was marked initially by the decline and fall of the huge Kushan 
Empire, the rule of the Hephthalites in Central Asia and their conflict with the Sasanians, 
while the close of the period coincides with the Arab conquest of the entire region. The 
area's development over these centuries was determined by two basic factors: the growth 
of feudal relations and outside invasions. 

There is as yet no irrefutable evidence that Khwarizm became part of the Kushan 
state. Opponents of this theory commonly cite as evidence of the country's independence 
the appearance of a local overstrike on Kushan coins in circulation in the territory of 
Khwarizm. 1 They further maintain that the wealth of archaeological material cannot prop- 
erly be taken as reflecting close links between Khwarizm and the Kushans, mainly because 
of the almost complete lack of any tangible vestige of Buddhism, which had been propa- 
gated in the lands conquered by the Kushans, particularly Tokharistan (Bactria). It should 
be noted, however, that not all the Kushan coins found in Khwarizm are overstruck; and 
investigation of the material and spiritual culture of the population in the early centuries 
of the Christian era points to some Indo-Buddhist connections. 2 However, influence in the 
field of art is not necessarily determined by conquest; it may emerge and spread through 
cultural and historical contact. Only if fresh data are amassed will it be possible to settle 
this vexed question. It is noteworthy that Toprak-kala (see also pages 213-17), an outstand- 
ing architectural group and grandiose dynastic centre, was built in the second century of the 
Christian era. This was undoubtedly a very significant event, and Tolstov is perhaps correct 

1 Masson, 1966, pp. 82-3; Vaynberg, 1977, pp. 87-9. 

2 Tolstov, 1948a, p. 201. 



ISBN 978-92-3-103211-0 Contents Uncertain early history 

in believing that the palace was built to mark the liberation of Khwarizm from dependence 
on the Kushans and in ascribing the appearance of a local overstrike on Kushan coins to 
that very period. 3 

In the early fourth century, Afrig came to power in Khwarizm after founding a new era, 
according to al-Biruni who lists 22 rulers of the dynasty founded by Afrig. 4 Numismatic 
evidence, third-century material from the archives of the Toprak-kala palace and inscrip- 
tions on the ossuaries from Tok-kala (seventh and eighth centuries) (Fig. 1), however, have 
significantly rectified this account. First, the notion of the Afrig era must be rejected, since 
it has been established beyond doubt that the Khwarizmian era, reflected in dates in the 
above-mentioned documentary material, began in the first century. 5 Furthermore, only cer- 
tain names of rulers from al-Biruni's list tally with those found on coins, 6 a discrepancy 
believed to be the result of his lack of reliable information on the pre-Islamic Khwarizmian 
dynasty. 7 It is thought, however, that the discrepancy will prove to be less important as 
more numismatic evidence is accumulated. 

Particular uncertainty surrounds the second half of the third century and the fourth cen- 
tury, a period which saw the appearance of many small copper coins and various seals. This 
seems to reflect a trend towards the political isolation of individual parts of Khwarizm. 8 
Between the last third of the third century (which corresponds to the reign of King Vaza- 
mar, possibly a usurper of the Khwarizmian throne) 9 and the end of the seventh century 
there are no satisfactorily dated coin series; indeed, there is a gap in the coinage. 10 It is 
important to note that the period in question corresponds to the time of the Sasanians' east- 
ward campaigns, Khwarizm being recorded as one of the countries they conquered. Thus, 
according to the Arab historian al-Tabari, Ardashir I seized Balkh, Merv and Khwarizm as 
far as the extreme limits of Khurasan. 11 However, the inscription of Shapur I on the Ka c be 
of Zoroaster at Naqsh-i Rustam, which lists the realms conquered by him, makes no men- 
tion of Khwarizm. 12 At the same time, numismatic and archaeological evidence suggests 
that Khwarizm was in some way dependent on the Sasanians. We may nevertheless ques- 
tion Henning's view that the country was totally subjugated for a considerable period. 13 

3 Tolstov, 19846, p. 16. 

4 Biruni, 1957, p. 48. 

5 Henning, 1965, p. 169; Livshits, 1968, p. 440; Vaynberg, 1977, pp. 77-80. 

6 Livshits, 1968, pp. 442-4; Vaynberg, 1977, pp. 80-2. 

7 Gudkova and Livshits, 1967, p. 10. 

8 Tolstov, 1948a, p. 183. 

9 See Tolstov, 1962, p. 225; Vaynberg, 1977, p. 97. 

10 Henning, 1965, p. 170; Livshits, 1968, p. 443. 

11 Noldeke, Tabari, 1973, pp. 17-18. 

12 Sprengling, 1953, pp. 7, 14; Lukonin, 1969, pp. 62, 126. 

13 Henning, 1965, pp. 169-70. 



ISBN 978-92-3-103211-0 


Social structure and administration 

FIG. 1. Tok-kala. Painted ossuary. (Photo: © Vladimir Terebenin.) 

We cannot rule out the possibility that Khwarizm was part of the state of the Turks in the 
sixth and seventh centuries, though relations between the Khwarizmians and the Turks are 
as yet obscure. 

The only events in Khwarizm about which the sources are comparatively clear are those 
that occurred at the time of the campaigns of conquest led by the Arab general Qutaiba b. 
Muslim (see below 'The Arab conquest'.). 

FIG. 2. Coin from Khwarizm (sixth century?). Bronze. 



ISBN 978-92-3-103211-0 


Social structure and administration 

Social structure and administration 

Between the fourth and the sixth century the cities of Khwarizm underwent a marked 
decline, with a cut-back in the irrigation network. This is usually blamed on a socio- 
economic depression, the scale of which is unclear, though it was more pronounced on 
the periphery of the country. It may have been at least partly the result of invasions by 
nomadic tribes at the time of the great migration of peoples, when the outlying areas were 
undoubtedly overrun. Furthermore, the country's internal situation must have been affected 
by the abovementioned events of political history. The seventh and eighth centuries, how- 
ever, saw some stability and even a measure of economic and cultural progress. 

We can only guess at the political and administrative organization of Khwarizm. There 
is some numismatic evidence of the existence of independent local rulers in the seventh 
and eighth centuries. For example, coins have come to light from the Kerder region of 
the lower valley of the Amu Darya (Oxus). That territory had been taken over by settlers 
from Khwarizm and by immigrants from the Syr Darya (Jaxartes) regions and its rulers 
succeeded in usurping the throne of the Khwarizmshah on several occasions. The 'king of 
Khamjird' seems to have ruled in northern Khwarizm. There is also a theory that Khamjird 
and Urgench were one and the same place. 14 

The monetary reform carried out in Khwarizm at some point in the fifth, sixth or seventh 
century led to the minting of new types of coins (Fig. 2) with a different value, and with the 
ideogram MR 'YMLK' (lord king) instead of the earlier MLK' (king). 15 What prompted the 
reform is unclear, but it may have reflected the desire of the Khwarizmshah to consolidate 

FIG. 3. Berkut-kala. Castle of fifth-seventh century (eastern Khwarizm). Aerial view. 

14 Vaynberg, 1977, p. 99. 

15 Ibid., pp. 59-65. 



ISBN 978-92-3-103211-0 Contents Social structure and administration 

his power over the other rulers. Al-Tabari mentions them when referring, in connection 
with the events of 711-712 (see pages 229 et seq.), to kings and dihqdm (lords). Clearly 
some hierarchical order or rule may be indicated by the coinage of the seventh century, 
which bears the Khwarizmian inscription hwt'w (lord) accompanying the portrait of the 
ruler, with the ideogram MR'Y MLK' on the reverse. Some authorities believe that the 
terms MLK', M7?Tand yw fiw in narrative sources and on Sogdian coinage of the seventh 
and eighth centuries also indicate three degrees in the local hierarchy. 16 In the Sogdian 
documents from Mount Mug, however, MR'Y (equivalent to yw fiw or xwt'w) is the title 
assumed both by Divashtich (the ruler of Panjikent) himself and by his underlings. 17 At 
the same time, the entire nobility of Iran of lower rank than the king bore the title of 
'lord'. 18 Nevertheless there can be little doubt about the general hierarchical structure of 
Khwarizmian society in the seventh and eighth centuries. Of particular significance is the 
archaeological evidence concerning Khwarizm's oasis settlements which sprang up along 
the major canals around the urban centres. 

The structure of the towns varied: the old towns founded in antiquity on the major 
trading routes continued to exist, still with a rectangular grid layout and fortified walls 
and towers; but new towns often sprang up by the walls of the castles of powerful feudal 
lords, reflecting the prevalent trend in the development of medieval towns. One example 
was the small town of Berkut-kala (Fig. 3): it was composed of two parts, one containing 
buildings occupied by craftsmen, around a market square, and the other entirely taken up 
with housing. 19 Another example is the vast complex of buildings (Fig. 4) at the foot of 
the fortress Ayaz-kala 2 (Fig. 5): here, in the fifth to the seventh century (and probably 
even before), a beautiful palace of the Khwarizmshahs emerged and a town developed 
near by (Fig. 6). Little is known about the appearance of the old towns. According to 
Arab authors, Madinat al-Fir, the capital of Khwarizm, was the country's largest and most 
strongly fortified town, the historian al- Baladhuri even comparing it to Samarkand. 20 It 
consisted of three parts surrounded by a moat: 21 the al-Fir citadel, the old town and the 
new part. In the tenth century, when the old town and the citadel had been almost destroyed 
by the Amu Darya, the new part grew into a town known at the time as Kath. 

Hazarasp was one of the country's main towns, according to Bal c ami. 22 Small-scale 
excavations there have revealed portions of strongly fortified walls and a large citadel 

16 Smirnova, 1963, p. 31. 

17 Livshits, 1962, p. 50. 

18 Harmatta, 1957, p. 303; Lukonin, 1961, pp. 16-17. 

19 Nerazik, 1966, p. 109. 

20 Al-Baladhuri, 1924, p. 188. 

21 Bolshakov, 1973, p. 171. 

22 Balcami, 1874, p. 176. 



ISBN 978-92-3-103211-0 


Social structure and administration 

■ .my.-- 

FIG. 4. Early Middle Ages complex near Ayaz-kala 2 (fifth-seventh century). 1: palace; 2: Ayaz- 
kala 2; 3: dwellings. 

may have existed in an angle of the town. It should be noted that Toprak-kala was a 
royal residence, a town specially built to serve a number of palaces, around which it lay 
(Figs. 7-9). The layout of the town, which covered 17.5 ha, was marked by great 
regularity throughout its existence (second to sixth century). The town of Kerder, the cen- 
tre of an independent domain, was laid out just as regularly as Toprak-kala, a fact which 
appears to reflect centralized urbanization. The Kerder settlements cannot be called towns 
in the full sense of the word, however. Surrounded by a primitive outer enclosure in the 
form of an embankment, and sometimes lacking outer walls, they arose spontaneously as a 
result of the settlement of the nomadic and semi-nomadic population around a central for- 
tification. The largest of them was Kuyuk-kala (41 ha), on whose territory have been found 
two ruined citadels belonging to different periods, vestiges of yurts and meagre traces of 
the handicrafts that were developed in the area, mainly by Khwarizmians. 



ISBN 978-92-3-103211-0 Contents Social structure and administration 


■*- *■* 

-* ■ 

FIG. 5. Ayaz-kala 2. Fortified complex. 

FIG. 6. Ayaz-kala 2. Palace near the fortified complex. 



ISBN 978-92-3-103211-0 


Social structure and administration 

FIG. 7. Toprak-kala. General view. (Photo: © Vladimir Terebenin.) 

FIG. 8. Toprak-kala. Reconstruction of the palace by Y. A. Rapoport. 

FIG. 9. Toprak-kala. Reconstruction of the interior of the palace by Y. A. Rapoport. 



ISBN 978-92-3-103211-0 Contents Social structure and administration 

In the agricultural oases of Khwarizm, groups of farmsteads lay along canals at whose 
outlets stood well-fortified castles. This reflects the hierarchical structure of a feudal soci- 
ety. The rural (and seemingly the urban) population lived in large patriarchal families span- 
ning three generations. Well-to-do families of the third century (whose houses were the 
BYT' of the Toprak-kala documents) employed many slaves, slavery then being patriarchal 
in character in the opinion of Livshits (Fig. 10). The house records use the term xrytk, 
meaning 'purchased'. 23 Here also, as in Sogdiana, there may have been slaves in debt 
bondage. 24 In inscriptions on ossuaries of the seventh and eighth centuries, where we find 
the same large kinship groups, the term 'Hunnanik' (son of a Hun) occurs; it is an indi- 
cation of yet another significant source of slaves - prisoners of war. These inscriptions 
include information on polygamy (which was common in wealthy families), 25 while the 
Toprak-kala documents also refer to concubines. 26 In the service of the wealthy there were 
also kedivars (dependent persons or clients), who had become landless peasant members 
of the community. 

Judging by excavations of dwelling-houses, ordinary farm labourers also lived in the 
large patriarchal families that made up rural communes. Some communities of this kind 
lived in the agricultural territory provided by an oasis, whose population may have attained 
4-5,000. Indirect evidence of the communal way of life can be deduced from such names of 
Khwarizmians as XwdnOacak ('possessed of good share'), fidySdrak ('holding a share') or 
('owning a garden'). 27 Each commune was a fairly closed world, based on a natural econ- 
omy. In the oases, cereals, vegetables, cultivated crops and fine-fibre cotton were grown. 
Home trades and crafts (potting, iron founding and smithery) almost satisfied the needs of 
the rural inhabitants. 28 After a short burst of activity in the third century, monetary circu- 
lation was not further developed until the seventh and eighth centuries, but the bulk of the 
coins still came from the castles. Silver and copper coins were in circulation; their face 
value and weight were approximately the same as other drachms of the time. 

With the accumulation of written and archaeological evidence concerning Khwarizm 
in the seventh and eighth centuries, it is becoming increasingly clear that the Sogdian 
and Khwarizmian societies developed along similar lines. It is important in this connec- 
tion that the three Sogdian social groups (?7/3), according to the documents from Mount 
Mug, included the tradespeople (yw'ky) in addition to the nobility ('rtkr) and the workmen 

23 Livshits, 1984, p. 267. 

24 Livshits, 1962, pp. 35-6. 

25 Gudkova and Livshits, 1967, p. 14. 

26 Livshits, 1984, p. 14. 

27 Ibid., pp. 269, 272. 

28 Nerazik, 1966, pp. 100-8. 



ISBN 978-92-3-103211-0 


Social structure and administration 

FIG. 10. Toprak-kala palace. Written document on wood giving the list of men living in the house 
of 'Harak'. (Room no. 90.) 

(k'rykr). 29 The merchant class also seems to have played an important part in the socio 
economic life of Khwarizm. Indeed, the Chinese chronicle the Tang shu [Annals of the 
T'ang Dynasty] makes special reference to Khwarizmian merchants journeying afar in ox- 
drawn carts, which seem to have been a feature of the country. 30 

Trading relations with the regions of the Aral Sea and the northwest Caspian area, 
together with the Volga and Ural regions, were extremely important for Khwarizm's econ- 
omy from the fourth to the eighth century. This trend in trading relations played a definite 
role in establishing Khwarizmian culture. Its influence was particularly marked in the peri- 
ods of the political predominance of Kerder (see above), and was felt, for example, in 
the interior layout of dwellings and the forms and ornamentation of ceramics. Overall, 
the culture of Khwarizm was the product of a complex interaction between, on the one 
hand, dominant, profoundly traditional local elements and, on the other, those brought by 
an immigrant population and arising from the country's historical and cultural contacts. 

29 Livshits, 1962, p. 37. 

30 Bichurin, 1950, pp. 315-16. 



ISBN 978-92-3-103211-0 


Art, architecture, religion and language 

While possessing a distinctive character and identity, Khwarizmian culture followed the 
course of development common to all artistic canons and forms elsewhere in the Early 
Middle Ages in Central Asia. 

Art, architecture, religion and language 

In the second half of the third century the unique complex of Toprak-kala was still in 
existence, embodying the highest achievements of the architects, painters and sculptors of 
Khwarizm (Figs. 1 1 and 12). Dynastic ceremonies were held in its many sanctuaries, which 
were sumptuously adorned with numerous paintings, coloured bas-reliefs and alabaster 
and clay sculpture in the round. The artistic traditions and the construction and architec- 
tural methods of the builders of the palace date back to antiquity. In the period from the 
fourth to the eighth century, private fortifications became widespread, and the castles of the 
aristocracy, with their mighty walls and towers, drawbridges, secret staircases and other 
structures, studded the countryside. That period saw the beginnings of the major types of 
popular dwellings and a number of public buildings of the Khwarizmian Middle Ages. At 
the same time, in view of the spread of polychrome subject paintings in the seventh and 
eighth centuries in other regions of Central Asia, the absence of monumental pictorial art 
in Khwarizm of the Early Middle Ages is striking. Although we know of nothing so far 

FIG. 11. Toprak-kala palace. Drawing from a mural painting. Woman with a thread. (Room no. 




ISBN 978-92-3-103211-0 


Art, architecture, religion and language 


FIG. 12. Toprak-kala palace. Drawing from the head of a statue in painted clay. (Room no. 36.) 

except painted ossuaries, the situation may change as more towns are excavated. 31 The 
applied arts are best represented in the form of engraving, toreutics and so on. 

On the brink of the Early Middle Ages in Khwarizm, traces of religious ideas that date 
back to remote antiquity, heathen agrarian cults and worship of the natural elements are 
frequently found. The kinds of deities which are apparent from the Khwarizmian names 
found in the Toprak-kala written documents - the water deity Vakhsh, the sun god Mithra, 
the fire deity and the god of the wind Vayu - could only have emerged in the context of 
ancient beliefs. 32 The traces of the Zoroastrian calendar in this material and the pantheon 
of deities that is constantly found in the art, together with the rituals of worship, make 
this system of beliefs comparable to the Avestan pantheon, 33 although there are marked 
differences. By the end of the third century, the earlier anthropomorphic representations of 
the deities had almost completely disappeared; fine plastic art was dying out, and statuary 
ossuaries were replaced by simple stone boxes. These changes are ascribed to the influence 
of orthodox Iranian Zoroastrianism, which banned idolatry. Finds of statuettes of idols, 

31 In fact, many fragments of subject paintings have been discovered during excavations of the palace of 
Ayaz-kala 2, but it is only after restoration that their meaning will become clear. 

32 Livshits, 1984, pp. 264-5, 272. 

33 Livshits, 1984, pp. 264-5, 272; Rapoport, 1971. 



ISBN 978-92-3-103211-0 Contents Art, architecture, religion and language 

however, particularly in strata from the fourth to the sixth century, testify to the vitality 
of heathen cults among the ordinary people of Khwarizm. The isolated terracotta figurines 
are quite different in form, their features bearing a strong resemblance to those of the 
stone idols of the steppe. They were introduced by the nomads living on the periphery of 
Khwarizm, perhaps at the time of the Turk campaigns. 

The ancient pantheon seems to have been preserved from the fourth to the eighth cen- 
tury, if sometimes in a barely perceptible symbolic form. For example, the four-armed god- 
dess on seals and silver dishes (Fig. 13) is seen as the ancient goddess Anahita, the goddess 
of fertility and the aqueous element, which brings felicity, 34 while the statuettes of horses 
and riders symbolize Mithra, whose cult was closely linked to the figure of Siyavush. This 
legendary ancestor of the Khwarizmian dynasty was also represented in the form of a 
horseman on the reverse of coins and on seals, and was the hero of a variety of ceremonies 
which were part of the great ritual cycle of New Year (Nowruz) celebrations. 35 

A complex interlacing of different religious beliefs, principally ancestor worship, can 
be observed in the burial ceremonies of the people, who placed the cleaned bones of their 
dead in ossuaries, which were then deposited in burial chambers or interred. Images on the 
ossuaries depicted both a real scene of mourning for the dead person and a number of ritual 
scenes showing the passing of the deceased, which were also part of the New Year cycle 
(Fig. 14). 36 

A major role in the system of religious beliefs was played by fire worship, which is 
extensively attested archaeologically. Fire sanctuaries in the form of small, single-chambered 
cupola-shaped towers have been discovered in the oases of east-bank Khwarizm (Fig. 15). 
One of these, dating back to the fourth century, is the oldest domed structure in Central 
Asia. Toprak-kala boasted a fire temple; and, as Rapoport rightly observes, the social strat- 
ification of the population was reflected in religious differentiations, for the townspeople 
worshipped in a temple unlike that used by the lords of the citadel and the palace. 37 The 
town's fire temple consisted of a small chain of chambers connected by a broad axial thor- 
oughfare. The temple of Hvarna (xvarenah oxfarn, the personification of power and mate- 
rial prosperity) was similar to the fire temple; richly decorated rams' horns symbolizing 
Hvarna were worshipped in it. 38 

No significant traces have been found of foreign religions in at that time. There is no 
indisputable evidence that Buddhism was commonly practised in Khwarizm, as distinct 

34 Tolstov, 1948a, p. 201. 

35 Ibid., pp. 200-3. 

36 Gudkova, 1964, pp. 95-100; Yagodin and Khojaiov, 1970, pp. 138-42. 

37 Rapoport, 1984, p. 296. 

38 Nerazik, 1982, pp. 52-4, 146. 



ISBN 978-92-3-103211-0 Contents Art, architecture, religion and language 

FIG. 13. Silver dish. (Photo: © Vladimir Terebenin.) 

FIG. 14. Tok-kala. Drawing from a painting on an ossuary in alabaster with a mourning scene (end 
seventh century). 

from other regions of Central Asia. On the other hand, the presence of Christian commu- 
nities in the country can be surmised. 39 A burial chamber apparently belonging to such a 

39 Tolstov, 1948a, pp. 227-8. 



ISBN 978-92-3-103211-0 


Art, architecture, religion and language 

FIG. 15. Oasis of Yakke-Parsan (eastern Khwarizm). Single-chambered cupola-shaped room with 
remains of an altar. 

community has been discovered in the necropolis of Gyaur-kala. 40 An independent ruler 
portrayed on a coin with a Nestorian cross on the crown was evidently a Christian. 41 

As archaeological work proceeds, increasing information is coming to light about the 
language of the people of Khwarizm in the period covering the third to the eighth century. 
Apart from inscriptions on coins and dishes, we now know of over 100 inscriptions on 
ossuaries, dozens of documents on skin and wood from Toprak-kala and 2 documents 
from Yakke-Parsan (eighth century). It has been established that the Khwarizmian language 
was related to the East Iranian group, while the writing was based on the Aramaic script. 
The inscriptions record dates of the Khwarizmian era, forms of greeting, proper names, 
terminology reflecting patrilineal kinship, and the names of the 12 months and of 18 days 
(out of 30), which may enable us to fill in some gaps in al-Biruni's data concerning the 
Khwarizmian calendar. 

40 Yagodin and Khojaiov, 1970, pp. 146-51. 

41 See Vaynberg, 1977, Table XXIX, 4. 



ISBN 978-92-3-103211-0 Contents The Khwarizmian calendar 

Part Two 


{P. G. Bulgakov) 

The Khwarizmian calendar 

Information on the history and culture of Khwarizm is given in the works of Abu Raihan 
al-Biruni, the learned Khwarizmian encyclopaedist who lived from 973 to 1048. His main 
source was the oral information of authorities on ancient traditions who were still alive 
when he was in Khwarizm, but he was also well versed in the historical literature which 
had been compiled by his time. In his work the Kitdb al-dthdr al-bdqiya [Chronology of 
Ancient Nations], known to modern scholars as the Chronology and written c. 1000-1003, 
al-Biruni gives information about the Khwarizmian calendar which is still the most detailed 
and comprehensive account we possess. 42 

Analysis of this information has established that the calendar was based on the so-called 
New Avestan Calendar, 43 but in time had taken on certain characteristics of its own. The 
Khwarizmian solar calendar consisted of 12 months, each of which had its own name 44 and 
consisted of 30 days, each of which also had a name. Five extra days were added to the end 
of the last (i.e. the 12th) month. These extra days did not have their own names, but were 
called by the same names as the first 5 days of the month. In this way, the Khwarizmian 
calendar consisted of 365 days. The additional 6 hours were not taken into account, and 
there was no leap year to correct this. The official names of the months and days given 
below are taken from Livshits' study. 45 

42 Biruni, 1957, pp. 62-3. 

43 See Bickerman's study of the Zoroastrian calendar (Bickerman, 1967, pp. 197-207). 

44 Al-Biruni gives two variations of the names of most months, one of which was the official name used by 
the Zoroastrian priesthood, and the other the name in everyday use. 

45 Livshits, 1970, pp. 167-9. More exact forms suggested by Livshits are given in brackets. 



ISBN 978-92-3-103211-0 


The Khwarizmian calendar 

I. r v rjn' 
II. V dw s t 

III. hrwd'd 

IV. jyry 
V. hm d' d 

VI. 'x s ry w ry 

1 . ry m zd (ry m zd) 

2. 'zmy n('h my n) 

3. 'rdwst 

4. 'xu}'w/ - }' 

5. '5 Z? n d' rmjy 

6. hrwd'd 
1 . hm d' d 

8. <5rfw 

9. Vw 

10. y' n' xn (y' b' x n) 

11. 'jc y r 

12. m' h (m' x) 

13. jyry 

14. y wil 

15. =8 


VII. 'w m ry ('m ry) 

VIII. y 'ft 'x ft (y '& ' x n) 

IX. Vw 

X. rymzd 

XL 'smn('hmn) 

XII. xitv/n 

16. /yy 

17. 'srwf 

18. r 5 n 

19. rw r/ft 

20. 're y n 

21. r' ft? 

22. w' 5 

23. =8,15 

24. dyny(SynorSyy) 

25. V/wxy 

26. 'st'S('st'd) 

27. 'sm'n 

28. z'?f=zflf?j 

29. m r s b n d 

30. Vnry 



ISBN 978-92-3-103211-0 Contents Khwarizmian eras 

As Livshits remarks, apart from occasional copyists' mistakes, al-Biruni's information 
is extremely accurate, and the names he gives for the months and days of the Khwarizmian 
calendar are in almost complete accordance with the inscriptions of the Tok-kala ossuaries. 46 

The Khwarizmian calendar came into being at the beginning of the Christian era and 
officially remained in use until Khwarizm was incorporated into the caliphate in the eighth 
century, when a system of chronology reckoning from the hegira was adopted at the time 
of the introduction of Islam. However, the people of Khwarizm continued to use the local 
calendar at least until the end of the tenth century. Its disadvantage was that, because there 
was no provision for the leap year, the beginning of the Khwarizmian months changed 
constantly from one time of the year to another. For this reason, according to al-Biruni, in 
960 the Khwarizmshah, Abu Sa c id Ahmad b. Muhammad b. Iraq, introduced a reform to 
make the beginning of the Khwarizmian months coincide with the fixed days of the Julian 
calendar. 47 

Khwarizmian eras 

In his Chronology, al-Biruni also mentions three Khwarizmian eras: the first beginning 
with the settlement of the country, which supposedly took place 980 years before the time 
of Alexander the Great (i.e. 1290 B.C.); 48 the second starting with the arrival of Siyavush in 
Khwarizm and the beginning of the reign of Kai Khusrau, 92 years after the settlement of 
the country; and the third beginning with the reign of the Khwarizmshah Afrig, who is said 
to have built a castle near Madinat al-Fir (Kath), the ancient capital of Khwarizm. 49 While 
it is generally agreed that the first two eras are legendary, certain scholars regard Afrig and 
his era as historical. 50 The genealogical list of the Khwarizmshahs of the Afrigid' dynasty 
given by al-Biruni would seem to support the authenticity of his information. For the period 
from 306 to the middle of the eighth century, this list is as follows: 51 

Afrig. Bagra. Sahhasak. Askajamuk (I). Azkajwar (I). Sahr (I). Shaush. Hamgari or Hangari. 
Buzgar. Arsamuh. Sahr (II). Sabri. Azkajwar (II). Askajamuk (II). Sawashfan. 52 

46 Ibid., p. 167. 

47 Biruni, 1957, pp. 262-3. 

48 It should be borne in mind that al-Biruni takes the year 310 B.C., not 3 12, as the beginning of the Seleucid 
era ('the era of Alexander'). Thus he dates the second year of the reign of Justinian I as the year 838 of the 
era of Alexander; and he counts the year 1191 of the era of Alexander as the year 267 from the hegira, and 
the year 1 194 of the era of Alexander as the year 270 from the hegira (Biruni, 1966, pp. 96, 128). 

49 Biruni, 1957, pp. 47-8. 

50 Tolstov, 19486, pp. 10-11, 191-2. 

51 Biruni, 1957, p. 48; Livshits, 1970, pp. 165-6, and his transliteration; Vaynberg, 1977, p. 83. 

52 Continuing this list, al-Biruni names a further seven Khwarizmshah^, the last of whom, Abu c Abdallah 
Muhammad, was killed in 995. 



ISBN 978-92-3-103211-0 Contents Religious beliefs 

So far, however, there is no confirmation of the use of a chronology based on the era of 
Afrig and attempts to read his name on Khwarizmian coins have proved unsuccessful. 53 
The archaeological material discovered in the last few decades - in particular, the archive 
of Toprak-kala and the Tok-kala ossuary inscriptions - in conjunction with the time scale 
of the Tok-kala culture have allowed us to establish that an official chronology existed in 
Khwarizm, but it was based on another age, for which we have no written sources. Accord- 
ing to Henning and Livshits, it began in the 10s or 20sa.d., and according to Vaynberg in 
the 40s or early 50s. 54 Thus, as was noted by Bartold, 55 al-Biruni's information cannot be 
regarded as reliable in its totality. 

The genealogical list of Khwarizmshahs, given by al-Biruni is also unreliable for the 
period up to the beginning of the eighth century. Of the early rulers, only the names of 
Azkajwar and Sawashfan are confirmed by Khwarizmian coins. The name of Arsamuh is 
also confirmed, but coins struck in his reign (no later than the end of the third century) do 
not coincide chronologically with his reign as given by al-Biruni. 56 Moreover, the names 
of the 13 Khwarizmshahs who reigned before Azkajwar as given on coins of Khwarizm are 
completely different from those given by al-Biruni. 57 The presence on most coins of the 
traditional tamgha (symbol or mark) for the period between the middle of the first century 
and the end of the eighth, and also the image of a horseman, give grounds for belief that 
there was a firmly established dynasty in Khwarizm during that period. However, there is 
no justification for calling it the Afrigid' dynasty. 

Religious beliefs 

The religion of ancient Khwarizm was a local version of Zoroastrianism mixed with pagan 
beliefs. In his Chronology, al-Biruni mentions traditions concerning the time and place 
of Zarathustra's ( Zoroaster's) activities. 58 Until recently little was known of al-Biruni's 
other work, the Al-Qanun al-Mas c udi [Canon of Mas c udi] (written over 30 years after 
the Chronology), because it had not been translated. In it al-Biruni reconsiders the date 
suggested in his Chronology for the beginning of the prophetic activities of Zarathustra 
- he now gives the year 276 before the Seleucid age ('the age of Alexander') or 1218 of 
the age of Yazdgird III, which was counted from the year a.d. 632. 59 Bearing in mind 

53 Livshits, 1970, p. 167. 

54 Henning, 1965, pp. 158-69; Livshits, 1970, pp. 163-5; Vaynberg, 1977, p. 79. 

55 Bartold, 1965, p. 545. 

56 The first third of the seventh century (Biruni, 1957, p. 48); Livshits, 1970, pp. 166-7; Vaynberg, 1977, 

57 Vaynberg, 1977, p. 81. 

58 Biruni, 1957, pp. 24, 205-6. 

59 Rlnini 1Q7-; nr. 11A 1/18 


Biruni, 1957, pp. 24, 205-6 
Biruni, 1973, pp. 114, 148. 



ISBN 978-92-3-103211-0 Contents Religious beliefs 

that al-Biruni believed that 310 B.C. was the beginning of the Seleucid age, not 312, and 
considering the dating according to the era of Yazdgird, we may conclude that, according 
to al-Biruni's latest information, Zarathustra began his activities not in 570 but in 586 B.C. 
If this is so, and if Zarathustra was then 42 years old, as tradition has it, then he was born 
108 years before the reign of Darius I. Thus al-Biruni's data are evidence against (and not 
for) attributing the activities of Zarathustra to the time of Darius I and his father. 

Al-Biruni informs us that certain Zoroastrian texts in the Khwarizmian language were 
destroyed as a result of the Arab conquest. 60 However, Zoroastrianism was not a legally 
established state religion in Khwarizm, as it was in Iran, and therefore did not follow strict 
canons. Both from al-Biruni's information and from archaeological evidence, it is clear 
that Zoroastrianism had a special character in Khwarizm, where it coexisted with survivals 
from earlier beliefs and local cults. Some of the manifestations of these were shared by 
pre- Islamic Sughd. 

Animistic notions such as the belief in jinns and in good and evil spirits survived in 
Khwarizm right up to al-Biruni's time (tenth to eleventh century). The cult of Vakhsh - 
the tutelary spirit of the element of water, especially of the Amu Darya - was a survival of 
early animism. As al-Biruni tells us, the Khwarizmian feast of Vakhsh was celebrated on 
the tenth day of the last (i.e. the twelfth) month of the year. 61 The link between the name 
of this spirit and the Amu Darya still survives in the name of the River Vakhsh, a tributary 
of the Panj, the upper reaches of the Amu Darya. There was a widespread belief in spirits, 
which were no longer associated with the elements or with plants and animals. Most were 
regarded as hostile to human beings. Smoke, steam and the smell of food were used to 
ward them off, as were certain types of ritual food prepared on set days, for example bread 
baked with fat on the first day of the seventh month of each year. 62 

The feasts of the New Year (the first day of the first month) and Chiri-Ruj (the thirteenth 
day of the seventh month), which coincided with the days of the vernal and autumnal 
equinoxes, were linked with the cult of nature, its death and revival. The beginning of the 
year marked the awakening and development of life-giving forces, while the autumn feast 
was its antithesis, and marked the time after which these forces faded away and died. 63 

The cult of the dead or of ancestors was observed with great respect in Khwarizm. 
According to al-Biruni, it was the Khwarizmian custom to place food in the burial cham- 
bers on the last five days of the twelfth month and five additional days of the New Year. 64 

60 Biruni, 1957, pp. 48, 63. 

61 Ibid., p. 258. 

62 Ibid., p. 257. 

63 Ibid., pp. 224, 234, 256, 257. 

64 Ibid., p. 258. 



ISBN 978-92-3-103211-0 Contents Religious beliefs 

This ritual of 'feeding' the ancestors was apparently linked with the belief that the fertility 
of the fields could be secured with the help of higher forces. 65 

In his Chronology, al-Biruni describes the rituals observed by the Sogdians in mem- 
ory of Siyavush, a legendary divine king who supposedly died as a result of being slan- 
dered at the height of his prosperity. It can be assumed that the same customs also existed 
in Khwarizm, as one of the legendary Khwarizmian ages was dated from the arrival of 
Siyavush in the country. 66 The commemoration of Siyavush, which was accompanied by 
sacrifices, was closely linked to the cult of the dead and to hopes of obtaining prosperity 
both on earth and after death. 67 It is possible that all the dead were honoured collectively in 
the form of Siyavush. 68 In the opinion of Tolstov, Siyavush was also venerated as the Cen- 
tral Asian god of dying and reviving vegetation. 69 The day of the ritual commemoration 
of the legendary Khwarizmian Queen Mina (who froze to death in warm weather), which 
was observed on the fifteenth day of the tenth month, was apparently also connected with 
the cult of the dead. 70 

Khwarizmian Zoroastrianism differed substantially from the canonical Iranian form in 
its burial rites. Whereas in Iran the bones of the dead were entombed in niches carved 
in rock or in vaulted burial chambers, 71 the Khwarizmians used ossuaries. According to 
Rapoport, the sources of the ossuary ritual should be sought not in Zoroastrian dogma, 
but in earlier beliefs. 72 The earliest statuary ossuaries, which were anthropomorphic and 
unacceptable to orthodox Zoroastrianism, were clearly a survival of ancient idolatry. The 
Khwarizmians continued to use them from the second century B.C. or slightly earlier until 
the third century a.d., when they were superseded by stone boxes as a result of the grow- 
ing influence of Zoroastrian dogma in Khwarizm. 73 Canonical Zoroastrianism forbade the 
mourning of the dead. In Khwarizm, however, as in Sogdiana, this ritual existed, as can be 
seen from the paintings on the Tok-kala ossuaries (see Fig. I). 74 

As for other religions, Buddhism probably never reached Khwarizm since there is no 
evidence of its having left substantial traces in the region. There was a community of 
Melkite Christians, whose clergy were under the jurisdiction of the Metropolitan of Merv. 

65 Rapoport, 1971, p. 115 

66 Bartold, 1971, p. 83. 
17 D'yakonov, 1951, pp. 34^3. 

D„„™,«,^ 1(V7 1 r. 51 

"' D yakonov, iyM,pp. 54-45. 

68 Rapoport, 1971, p. 83. 

69 Tolstov, 1948a, pp. 202-4; 1948/?, pp. 83-7; Gafurov, 1972, pp. 

70 Biruni, 1957, p. 257. 

71 Herzfeld, 1935, p. 39; 1941, pp. 217-18. 

72 Racooort. 1971. do. 5. 18. 32. 


Rapoport, 1971, pp. 5, 18, 32 
73 Ibid., pp. 120-1. 

Gudkova, 1964, pp. 95-102. 




ISBN 978-92-3-10321 1-0 Contents The Arab conquest 

Among the Melkite feast days celebrated in Khwarizm, al-Biruni mentions the feast of 
roses, which were brought to church on 4 May (Ayyar) every year. 75 

Al-Biruni's Chronology is the only source of information about the existence of certain 
secular holidays in pre-Islamic Khwarizm. In connection with the agricultural calendar, 
the first day of the third month was observed as the beginning of the sowing season for 
sesame and some other crops. 76 The fifth day of the fourth month was also kept as a feast; 
the Khwarizmians counted seventy days from that date, and then began sowing the winter 
wheat. 77 The feast celebrated on the first day of the sixth month had its roots in their former 
nomadic way of life: according to tradition, at that time (which in the past had coincided 
with the beginning of the cold season), the kings of Khwarizm used to leave their summer 
quarters and go out onto the steppe to protect their lands from nomadic raids. 78 

In ancient and early medieval Khwarizm, astronomical observations were made for both 
religious and practical purposes. Archaeological data give grounds for believing that as 
early as the second and third centuries B.C., a great temple stood on the site of the ruins of 
Koy-Krylgan-kala. It was built where an ancient mausoleum had stood, and was used both 
as a burial place and as a centre of the cult of the dead and of the stars. 79 Observations 
were concentrated mainly on the sun and the moon, both because they were the most 
highly venerated in Zoroastrianism and because of their role in measuring time. The ancient 
Khwarizmians were familiar with eclipses of the sun and moon (the 'houses' of the moon), 
and were therefore able to determine and correct the times of the seasons and thus the 
calendar system as a whole. In his Chronology, al-Biruni gives the Khwarizmian names 
for all the 12 signs of the Zodiac, the 28 'houses' of the moon (the groups of stars in 
which the moon 'stands' on each of the 28 days of the lunar month) and also the names of 
the Sun, Moon, Mars, Venus, Jupiter and Mercury. 80 He observes that the Khwarizmians 
knew more about many constellations than the pre-Islamic Arabs did, and gives a number 
of examples. 81 

The Arab conquest 

The Arab conquest of Khwarizm and the country's subsequent conversion to Islam pro- 
voked a crisis in the indigenous culture. The first attempt by the Arabs to conquer Khwarizm 

Biruni, 1957, pp. 318, 326. 
Biruni, 1957, p. 256. 


77 Ibid., pp. 256-7. 

±uiu., pp. wu-J. 

78 Ibid., p. 257. 

79 'Koy-Krylgan-Kala', 1967, pp. 235-6, 253-64. 

80 Biruni, 1957, pp. 187-8, 261. 

81 Ibid., p. 259. 



ISBN 978-92-3-10321 1-0 Contents The Arab conquest 

came at the very end of the seventh century. Umayya b. c Abdallah, the Arab governor of 
Khurasan (693-697), after capturing the capital city of Kath (which he held for a time), 
forced the Khwarizmshah to sign a treaty recognizing the power of the caliphate. Imme- 
diately after the departure of the Arab forces, however, this treaty was abrogated. 82 The 
attempts of Yazid b. al-Muhallab, another governor of Khurasan (702-705), to take pos- 
session of Khwarizm were equally unsuccessful 83 and it was not until 711-712 that his 
successor, Qutaiba b. Muslim, succeeded in doing so, by exploiting the civil war that was 
raging in the country. 

According to al-Baladhuri, Khurrazad, the younger brother of the reigning 
Khwarizmshah, raised a revolt against his brother, became de facto ruler, took the law 
into his own hands and robbed the local nobles. As the Khwarizmshah could not with- 
stand Khurrazad, he sent messengers secretly to Qutaiba b. Muslim, who was in Merv. 
The Khwarizmshah agreed to pay tribute to Qutaiba and to recognize the power of the 
caliphate on condition that he remained lord of Khwarizm and that Qutaiba deliver him 
from Khurrazad's oppression. Qutaiba agreed and sent his brother, c Abd al-Rahman b. 
Muslim, against Khurrazad. He was victorious in the battle and Khurrazad was killed. 
c Abd al-Rahman publicly executed the 4,000 prisoners he took. 84 

Al-Tabari, and later Bal c ami and Ibn al-Athir, tell us that in addition to Khurrazad the 
Khwarizmshah had another sworn enemy, a certain King Khamjird. 85 According to the 
combined information provided by these historians, the sequence of events is as follows. 
In the year 93 a.h. (a.d. 712), a Khwarizmshah by the name of Jigan (or Chigan) 86 was 
faced with his younger brother Khurrazad's open flouting of his authority and with an even 
more powerful enemy, King Khamjird. Without informing anyone, he turned for help to 
Qutaiba b. Muslim, who was then in Merv. As a sign that he recognized the power of 
the caliphate, the Khwarizmshah sent Qutaiba the golden keys to the three main cities of 

82 Al-Baladhuri, 1866, p. 426. 

83 Al-Baladhuri, 1866, p. 417; al-Tabari, 1881-89, pp. 1142-3. 

84 Al-Baladhuri, 1866, pp. 420-1. 

85 The term malik Khamjird is understood by some scholars (Tolstov, Vyazigin) as 'King Khamdzhird' 
(Khamdzhard) and by others (Bartold, Vaynberg) as 'the king of Khamjird'. The title The Conquest of 
Khamjird in Ibn al-Athir's Chronicle suggests that Khamjird was a toponym. Vaynberg, basing his opin- 
ion on one of the editions of Bal'ami's work in which Gurganj appears instead of Khamjird, considers that 
it is the same town. The fact that Khamjird is not mentioned in other sources can be explained, according to 
Vaynberg's hypothesis, by the fact that the town's old name is replaced by its new one (Ibn al-Athir, 1301 a.h., 
p. 273; Istorya Turkmenskoy SSR, 1957, pp. 163-4; Tolstov, 1948b, p. 225; Bartold, 1965, p. 546; Vaynberg, 
1977, pp. 98-9). 

86 Only Bal c ami gives this name and in the manuscripts of his work kept in the Institute of Oriental Studies 
of the Uzbek Academy of Sciences it is spelt in different ways. As well as the most common form, there are 
others which are graphically close to them (BaFami, MS 33, p. 359 a; MS 4226, p. 371 b; MS 6095, pp. 368 
b-369 a; MS 7466, pp. 273 b- 274 b; MS 1 1273, p. 459 b). 



ISBN 978-92-3-10321 1-0 Contents The Arab conquest 

Khwarizm. Qutaiba set out with his forces from Merv, ostensibly on a campaign against 
Sogdiana. Rumours to this effect were also spread by the Khwarizmshah to put those who 
favoured war with the Arabs off their guard. When Qutaiba unexpectedly appeared with 
his army in Hazarasp, the Khwarizmshah advised his entourage not to resist, in view of 
the obvious superiority of the Arab forces. The Khwarizmshah concluded a treaty with 
Qutaiba according to which he gave him 10,000 head of livestock, money and other prop- 
erty, on condition that Qutaiba would advance against King Khamjird. Qutaiba accepted 
these conditions. His brother, c Abd al- Rahman b. Muslim, won a victory over the ruler of 
Khamjird, and the latter was killed. c Abd al-Rahman took 4,000 prisoners, whom Qutaiba 
ordered to be executed. Then Qutaiba captured Khurrazad and his henchmen and handed 
them over to the Khwarizmshah, who had them executed. 87 

Having achieved his aims in Khwarizm, Qutaiba besieged and took Samarkand in the 
same year, 93 a.h. (a.d. 712); during this campaign the people of Khwarizm revolted and 
killed the Khwarizmshah, who had betrayed their country. When Qutaiba returned from 
Sogdiana, he dismissed Ilyas b. c Abdallah b. c Amr, his commissioner in Khwarizm, for 
failing to take action and sent c Abdallah b. Muslim to replace him. Although the latter 
maintained his authority for a time, revolt soon broke out again and was not suppressed 
until al-Mughir b. c Abdallah, another of Qutaiba's generals, was sent there with his forces. 
It was then, according to al-Biruni, that Khwarizm was sacked by barbarians; as a result, 
many objects of cultural value were destroyed, including Khwarizmian manuscripts. 88 

It is reasonable to assume that there are considerable gaps in the medieval histori- 
ans' account of the violent events which took place in Khwarizm in 712. Coins of the 
Khwarizmshah Azkajwar II have been found overstruck by the types of a certain Khusrau, 
who only reigned for a short time. 89 The most likely explanation is that Khusrau was put in 
power by the Khwarizmians when they revolted and overthrew Jigan. Then, after Qutaiba 
had suppressed the rising, he put a member of the old ruling dynasty, Askajamuk II, the 
son of Azkajwar II, on the throne. 90 

The absence of the name of Jigan (Chigan) on Khwarizmian coins and in the genealog- 
ical list of Khwarizmshah^, given by al-Biruni has given rise to two hypotheses. Accord- 
ing to the first, Jigan was a usurper, who unsuccessfully tried to consolidate his power 
with the help of the Arab conquerors, 91 but who ruled for such a short period that he 
did not have time to issue his own coinage. According to the second hypothesis, the 
one most widely held by scholars, Jigan is an etymologically obscure nickname of the 

87 Al-Tabari, 1879-89, pp. 1236-9; Bal'ami, MS 4226, pp. 371 b-372 a; Ibn al-Athir, 1301a.h., pp. 273-4. 

88 Biruni, 1957, pp. 48, 63. 

89 Henning, 1965, pp. 168, 175; Vaynberg, 1977, pp. 78-9. 

90 Biruni, 1957, p. 48; Gudkova and Livshits, 1967, p. 6; Livshits, 1970, p. 164. 

91 Gudkova, 1964, pp. 120-1. 



ISBN 978-92-3-10321 1-0 Contents The Arab conquest 


Azkajwar II, the father of Askajamuk II. 92 The available data do not yet allow the issue to 

be settled. 

Early Arab rule in Khwarizm was unstable. In 110 a.h. (a.d. 728) a rising supported 
by neighbouring Turk tribes broke out in the north of the country, in the town of Kerder 
and the area around it, but it was suppressed in the same year. 93 From then on, the political 
situation was somewhat more stable. Members of the ancient dynasty continued to reign 
over Khwarizm as vassals of the caliphate with limited rights until 995, when the last of 
them was executed by the amir of Urgench, who united the entire country under his rule. 

92 Gudkova and Livshits, 1967, p. 6; Vaynberg, 1977, pp. 81, 91-3. Rtveladze has given his reasons for 
objecting to Vaynberg's theory that there was a dynastic link between Jigan and Chaganiyan (Rtveladze, 
1980, pp. 51-8). 

93 Al-Tabari, 1881-89, p. 1525. 



ISBN 978-92-3-10321 1-0 Contents SUGHD AND ADJACENT REGIONS 


B. I. Marshak and N. N. Negmatov 



Economic, cultural and social life 242 

Art and architecture 244 

Religious life 249 

Scripts, epics and literary sources 255 


Ustrushana 262 

Ferghana 275 

Chachandllak 278 

Part One 


(R /. Marshak) 

During the third to the eighth century Sughd (Sogdiana) included the basins of the 
rivers Zerafshan and Kashka Darya. The name 'Sughd' was frequently applied only to 

See Map 5. 



ISBN 978-92-3-10321 1-0 Contents SUGHD AND ADJACENT REGIONS 

the area near Samarkand - Samarkandian Sughd - but sometimes it was extended to the 
whole area where the Sogdian language was predominant, which in the seventh century 
included regions to its north-east (Ustrushana, Chach - the modern Tashkent - and west- 
ern Semirechye; see Part Two). From the third to the eighth century Sughd, which had 
originally lagged behind its neighbours to the south and west, became one of the most 
advanced countries and the leader of all Transoxania. It was neither a powerful state itself 
nor firmly subjected to any of the neighbouring empires. From the second or first century 
B.C. each district had developed independently, maintaining ancient community traditions. 
Private individuals such as merchants, missionaries and mercenary soldiers were extremely 
active and penetrated into distant lands. Thus political isolation did not entail cultural 

Other peoples knew the Sogdians mainly as silk merchants, but the basis of the Sogdian 
economy was agriculture on artificially irrigated land. From the very beginning, the Silk 
Route was controlled by Sogdian merchants, but in the fifth century their domestic trade 
and monetary relations were still at the stage of 'Barbarian imitations'. 1 Century after 
century each of the main areas of Sughd minted coins that can be traced back to the coinage 
of various Hellenistic rulers of the third to the second century B.C. Moreover, the coins of 
Samarkandian Sughd had become extremely debased by the fifth century; the image of an 
archer had been reduced to a mere outline and the weight of the coin was considerably 
reduced. Although Sughd was a neighbour of the Kushan Empire and was invaded by 
Iranian forces during the third century, it was not incorporated into these states with highly 
developed administrative systems. In the sixth century the minting of coins with the image 
of an archer, which had continued for many centuries, ceased; this marked the end of the 
stage of 'Barbarian imitations' and the beginning of a new stage in the development of 
trading and monetary relations. 2 

In ancient times large numbers of nomadic or semi-nomadic herdsmen lived around 
Sogdiana. Some of their burial places have been excavated near the borders of oases; the 
ceramics found in the graves are of Sogdian workmanship, showing the crosscultural influ- 
ence. At some time between the third and the fifth century (the exact date has not yet been 
established) these burial places fell into disuse, settlements were destroyed and sometimes 
even deserted, and the craft traditions, as revealed by the forms of the wheel-turned pottery, 
changed. In the fifth-century strata in the Kashka Darya valley there are large quantities of 
handmade ceramic articles, together with turned pottery. People with a tradition of hand- 
made pottery, characteristic only of settled peoples and found to this day among the Tajik 

1 Zeimal, 1983, pp. 69-76, 269-76. 

2 Zeimal, 1983, p. 234. 



ISBN 978-92-3-103211-0 



hillsmen, migrated from the backward outskirts to the partially abandoned fertile land. 
Around the fourth century what is called the second wall, enclosing an area of 66 ha, was 
built inside the ancient wall of the capital city, Samarkand - there were clearly not enough 
people to defend the old wall, which was almost 6 km long. 

Between the third and the seventh century there were no prolonged periods of decline. 
New settlements appeared in both the third and the fourth centuries, and during the fifth 
century whole towns were built, including Panjikent, 60 km east of Samarkand (Fig. 1). The 
evidence of local crises in the third and (especially) the fourth century can be explained by 
the migration of new groups of nomads who appeared on the borders of Iran in the middle 
of the fourth century, and in the following century began a long struggle with Iran for 
Tokharistan. It was probably the arrival of the Sasanians that drove the local semi-nomadic 
herdsmen away from the land on which they had settled. 

Four generations before the beginning of relations between the Wei dynasty and for- 
eigners from the West, i.e. in the second half of the fourth century, the Huns killed the ruler 
of Sogdiana and seized possession of his lands. 3 The Sogdians called this people xwn. Were 
these the same Huns who overran northern China in the fourth century or were they Chion- 
ites, whom the Indians called Hunas? There was a difference in the physical appearance of 
the two peoples, although some of the Huns were probably included among the Chionites. 4 
The Huns who appeared in Sogdiana were probably not the same people as the Chionites, 

FIG. 1. Panjikent. Excavations of the palace on the citadel. (Photo: © Vladimir Terebenin.) 

3 Enoki, 1955. 

4 Czegledy, 1980, pp. 213-17. 



ISBN 978-92-3-10321 1-0 Contents SUGHD AND ADJACENT REGIONS 

but even if they were, no Chionite Empire encompassing Sughd and Tokharistan existed in 
the fifth century. It is believed that the Chionites (the sources mention Turks) attacked Iran 
from the direction of Sughd. 5 This was probably not, however, a real historical event from 
the fifth century but an anachronistic episode from the Romance of Bahrain Gur inspired 
by the later victory of the Sasanian general Bahram Chobin over the Turks. It is not impos- 
sible that in the fifth century Sughd was ruled by a dynasty of nomadic origin, although 
there is no evidence of the activities of nomads in the country. The increased strength of 
the nomads in Tokharistan in the fourth and fifth centuries may have been connected with 
their departure from Sughd. 

Sughd came under the rule of the Hephthalites in c. 509, from which time Hephthalite 
'embassies' from Samarkand (consisting essentially of trading caravans) to China were 
known. 6 The Hephthalites seem to have come to Sughd from the south after victory over 
Sasanian Iran. For the first time since the Seleucids, Sughd came under the power of an 
organized state with an army rather than the Sogdian rulers' retinues, and this army was 
considerably stronger than the local militia. 

In Tokharistan, where tribes of Hephthalites lived, Chinese travellers report seeing 
nomads with archaic customs, but Byzantine authors describe the Hephthalites as an urban 
population with a highly organized state. Menander's information about the urban Hep- 
hthalites goes back to the conversation of Sogdian envoys of the Turk kaghan with the 
Byzantine emperor and hence refers to the Hephthalites in Sughd. The town of Panjikent 
grew during the Hephthalite period, its fortifications were strengthened and temples were 
rebuilt, although the arrival of the Hephthalites is thought to have been accompanied by 
ravages (Fig. 2). 

At the end of the Hephthalite period or the beginning of the Turkic period (during the 
sixth century), a ruler by the name of Abrui held sway over the oasis of Bukhara. This tyrant 
drove the nobles and the rich to emigrate to Semirechye. 7 Contrary to Tolstov's opinion, 
there is no mention in the sources of the people's struggle against the aristocracy; on the 
contrary, the poor who remained under Abrui 's rule begged the emigrants for help. Abrui 
was finally defeated by Turk forces who came at the request of the people of Bukhara. 
(Abrui himself was not of Turk origin: although he and the Bukharans are described as 
coming 'from Turkestan', this refers to the earliest days of the settlement of the oasis of 
Bukhara.) Legend tells of a prince's struggle against a community of citizens headed by 
nobles. Such a struggle might have been all the more bitter because of the rapid growth of 

5 Marquart, 1901, pp. 50-1. 

6 Enoki, 1959. 

7 Marquart, 1901, p. 309; Tolstov, 1948, pp. 248 et seq.; Gafurov, 1972, pp. 223 et seq. 



ISBN 978-92-3-103211-0 



FIG. 2. Panjikent. Plan of the city (without the citadel). (Drawing by B. I. Marshak.) 

the towns in the fifth to the seventh century, as revealed by the archaeological evidence. 
Paikent, the residence of Abrui, became a free 'merchant city' in the seventh to the eighth 

Typically, in the legend of Abrui, justice is re-established due to the arrival of the Turks. 
When the inheritance of the Hephthalites was divided up between the Turks and the Sasa- 
nians (who had defeated the Hephthalites in the 560s), Sughd fell to the Turks, although 



ISBN 978-92-3-103211-0 Contents Economic, cultural and social life 

their forces probably left the country after the victory. They came again in the 580s, when 
the war with Iran began. 8 After establishing the kaghanate over a vast territory stretching 
from the Black Sea to the Chinese border, the nomadic Turk kagham recruited Sogdian 
civil servants to run it. The Sogdian colonization of Semirechye and the Sogdian caravan 
trade were of benefit not only to the Sogdians but also to the Turks. The Turk state aspired 
to make the roads safe and gave its backing to the Sogdian diplomats' trade negotiations. 
The Sogdian language, which had become the lingua franca of the Silk Route long before 
the sixth century, became the official language of the kaghans' administration in the second 
half of the sixth century. 9 During the first third of the seventh century, after the division into 
Eastern and Western Kaghanates, Turk influence increased, but the kaghan and the king 
of Samarkand were now more like equal allies than sovereign and subject. The kaghan's 
daughter married the king of Samarkand. 

In the middle of the seventh century, after the fall of the Western Kaghanate, the Sogdian 
states gained de facto independence, although formally recognizing the sovereignty of the 
T'ang dynasty. In the eighth century, this sovereignty proved to be purely nominal, because 
China gave no real support against the Arab invaders. The alliance with the Turk states was 
unstable, with the Turkic nobles frequently looting or seizing Sogdian territories; as early 
as the end of the seventh century the principality of Panjikent had a Turk ruler, Chikin Chur 

In the second half of the seventh century, after conquering Iran, the Arabs advanced 
on Sughd. During the first few decades of the eighth century Arab garrisons were estab- 
lished in Bukhara and Samarkand and the local rulers submitted. The suppression of local 
uprisings in 720-730 led to mass emigration. In 739 the Arabs concluded a treaty with the 
Sogdians, many of whom returned home and, as the excavations in Panjikent show, tried to 
re-establish their former way of life. Mass conversions to Islam began in the 750s, but the 
process of Islamization and the gradual waning of the power of local rulers took several 
more decades. 

Economic, cultural and social life 

The seventh century and the first half of the eighth were the economic and cultural heyday 
of pre-Islamic Sughd. However, evidence from the documents of Dunhuang reveals the 
existence of large colonies of Sogdian merchants in Chinese towns 10 as early as the begin- 
ning of the fourth century (if not the end of the second) and, judging by the Karakorum 

8 Gafurov, 1972, pp. 217-21. 

9 Klyashtorny and Livshits, 1971; 1972. 
10 Henning, 1948; Harmatta, 1971. 



ISBN 978-92-3-103211-0 Contents Economic, cultural and social life 

rock inscriptions, Sogdian merchants predominated at that time on the southern routes from 
Central Asia to the Indus valley. 11 The general upsurge in the domestic economy allowed 
the profits from foreign trade to enrich not only the emigrant merchants but the population 
of Sughd itself. Silk weaving began in the country in the late sixth and early seventh cen- 
turies and by the following century Sogdian silks were already playing a major role on the 
Silk Route. 12 From the sixth to the seventh century the 'Fur Route' to north-west Europe 
was also in the hands of Sogdian and Khwarizmian merchants - as revealed in the Sogdian 
and Khwarizmian owners' inscriptions on Byzantine and Iranian silver vessels found in the 
north, where they were imported in exchange for furs. 13 In Sughd itself the comparatively 
primitive bowls of the sixth century had, by the seventh century, given way to magnificent 
artistic vessels and sculptures fashioned in silver. 14 

The changes in life-style were reflected even in such a 'democratic' craft as ceramics. 
The second half of the seventh century and the beginning of the eighth saw a complete 
change in the design of table crockery, whose shapes and ornamentation began to copy 
those of the nobles' silver vessels. 

The new stage in the development of trading and monetary relations was associated 
with the wide circulation in Sogdiana of a cast bronze coin with a square hole in the middle 
(Fig. 3). 15 The coins of Samarkand, Panjikent, Paikent and certain other centres are well 
known. Silver drachms were minted in imitation of the Iranian drachms of Bahram V 


FIG. 3. Bronze coins from Panjikent. (Photo: © Vladimir Terebenin.) 

11 Humbach, 1980. 

12 Yerusalimskaya, 1972. 

13 Livshits and Lukonin, 1964. 

14 Marshak, 1971. 

15 Smirnova, 1981. 



ISBN 978-92-3-103211-0 


Art and architecture 

(420-438). Their inscription normally included the title of the ruler of Bukhara, although 
in the eighth century they were issued in other centres as well. 16 

Art and architecture 

There was architectural progress in towns and settlements such as Samarkand (the site of 
Afrasiab) (Fig. 4), Panjikent and Varakhsha (near Bukhara) (Figs. 5 and 6). In the seventh 
century Samarkand again covered the whole plateau of Afrasiab, an area of 219 ha. Other 
Sogdian towns were much smaller. The area of Bukhara (without the citadel) was 34 ha, 
and that of Panjikent (also without the citadel) 13.5 ha. The buildings within the city walls 
have been best studied in Panjikent. 17 In the fifth century the residential quarters were 

FIG. 4. Afrasiab. View from the old city towards Samarkand. (Photo: © Vladimir Terebenin.) 

16 Davidovich, 1979, pp. 105-6. 

17 Belenitskiy et al., 1981. 



ISBN 978-92-3-103211-0 


Art and architecture 

FIG. 5. Varakhsha. General view. (Photo: © Vladimir Terebenin.) 

FIG. 6. Varakhsha.Corridor. (Photo: © Vladimir Terebenin.) 

composed of detached houses, but over the sixth, seventh and early eighth centuries the 
whole town was built up with uninterrupted terraces in each quarter. The houses were 
of compressed clay (loess) and mud-brick and roofed with mud-brick vaults or wooden 
structures, which were then plastered with clay. In the sixth century upper storeys began to 



ISBN 978-92-3-103211-0 


Art and architecture 

FIG. 7. Panjikent. Palace of Divashtich. Drawing by A. Gurevich. © B. I. Marshak. 

appear, and by the eighth century some buildings even had three storeys. The houses of the 
late seventh century and the first half of the eighth were more spacious, taller and in every 
way superior to the earlier dwellings. 

Panjikent reached the height of its prosperity during the first quarter of the eighth cen- 
tury, when its ruler Divashtich claimed power over all (Fig. 7). Private houses with murals 
had started to be built in Panjikent as early as the sixth century; in the early eighth cen- 
tury, one house in three had murals. In their ornate architecture and decoration, the rich 
town houses resemble the royal palaces discovered in Panjikent, Varakhsha 18 and, per- 
haps, Samarkand. 19 Although the palaces contain several large state apartments, they are 
basically very similar to the houses of wealthy townsfolk. This is because of the particu- 
lar Sogdian social structure, in which an important role was played by urban communities 
with their own officials and revenues. During the seventh and eighth centuries, the rulers 
(who enjoyed no absolute or despotic powers over the city-states) were frequently elected 
by the notables. 

The shopkeepers and craftsmen lived in two-storey houses with several rooms, but 
mainly did business and worked in rented shops and workshops located in wealthier areas. 
Coins have often been discovered in these shops. 

18 Shishkin, 1963. 

19 Al'baum, 1975. 



ISBN 978-92-3-103211-0 Contents Art and architecture 

Fortified homes belonging to the country aristocracy had existed in Sughd since Hel- 
lenistic times, becoming particularly numerous in the Early Middle Ages. In the seventh 
and eighth centuries the chambers in castles were very similar to those in the town houses 
of the wealthy, whereas the peasants' homes were unlike those in towns and resembled 
nineteenth-century Tajik peasant houses. 

Sogdian temples are known from the site of Er-kurgan in south (fourth-sixth century) 
and from two temples discovered in Panjikent. Of similar design and built at the same 
time in the fifth century, the Panjikent temples were rebuilt several times and continued 
in use until after 720; although they have been subjected to more detailed study than the 
temple at Er-kurgan, it is hard to know to which religion they belonged as the main images 
worshipped there have not survived. The architectural plan of both temples was based on a 
road leading east to west and passing through two rectangular courtyards with colonnaded 
entrance porticoes; from here a narrow ramp led up to the platform of the main building, 
which also had a portico. A four-columned hall without an east wall opened on to it. A 
door in the west wall of the hall led to a rectangular cell. A gallery ran round the hall and 
the cell on three sides. During the late fifth and early sixth centuries one of the temples 
had a special chamber for the sacred fire, but no such chambers of earlier or later date 
have been found. The temples were similar (but not identical) in plan to Kushan and even 
Graeco-Bactrian examples. 

The murals in Sogdian towns - not so much those in the temples but, rather, in seventh- 
and eighth-century private houses - depict daily life. In the ceremonial hall of a house, 
opposite the entrance, there was usually a large image of a god (or a more elaborate com- 
position on a sacred theme) and small figures of Sogdians before a fire altar (Fig. 8). 
Every house owner had his own divine patron (or patrons). Sometimes other gods were 
also depicted, but on less important parts of the walls. To the sides of the religious scene 
ran smaller friezes depicting a banquet (Figs. 9 and 10), a hunt (Fig. 1 1), a ceremonial rite 
or - quite frequently - episodes from an epic. At the bottom of the wall ran an ornamen- 
tal border or a frieze of small pictures: animals in motion or a series of small rectangular 
panels illustrating tales, parables, anecdotes, and so on (Fig. 12). The elaborate pictorial 
scheme was complemented by carved wooden statues and reliefs (Figs. 13 and 14), with 
figures of gods, hunting scenes, and so on, adorning the elaborate wooden ceiling. 20 

This hierarchy of subjects was typical of the main halls of Panjikent, but was not so 
rigidly followed in Varakhsha (Fig. 15) and in Afrasiab (nor, indeed, in some houses in 
Panjikent itself or its temples). Palace murals include subjects from Sogdian history, the 
reception of ambassadors (Afrasiab) (Figs. 16 and 17), a coronation and the Arab siege of 

20 Azarpay, 1981. 



ISBN 978-92-3-103211-0 


Art and architecture 

FIG. 8. Panjikent. Ceremonial hall in a private house. Drawing by B. I. Marshak. 

the town (Panjikent). Particularly expressive are murals with figures picked out in ochre 
on an ultramarine blue background. A similar colour scheme is found in the murals of 
Iran and Tokharistan, which share certain other details, attitudes of figures and types of 
composition. There are many Indian traits in the depiction of gods. There are also pictures 
of Chinese people, drawn with a knowledge of Chinese iconography. Motifs of foreign 
origin (Byzantine, Iranian, Turk and Chinese) can also be traced in the metalwork (Figs. 
18-20). These varied influences are explained by Sughd's role as intermediary. Sogdian 
artists were familiar with the achievements of other schools of art, but developed their 
own original style, distinguished by its narrative content, dynamism and love of contrast. 
Sogdian art had a strong influence on that of many countries, in particular on the toreutics 
of the steppe peoples (Turks, Khazars, nomadic Magyars) and of T'ang China. 

Archaeological discoveries show that Sogdian artists had been faithful in their depic- 
tions of architecture, weapons and costume. Fifth-century Sogdian costume was similar to 
that in the Kushan Empire; during the sixth century the influence of Hephthalite Tokharis- 
tan (and through it that of Sasanian fashions) was noticeable. In the seventh and eighth 



ISBN 978-92-3-103211-0 Contents Religious life 

FIG. 9. Panjikent. Mural painting. Clay on plaster. (Photo: © Vladimir Terebenin.) 

centuries similarities with Turkic costume appeared. Belts decorated with gold plaques 
were the mark of noble rank. Military clothing and equipment (Fig. 21) and, to some 
degree, vessels used in banquets also showed Turkic influence. Sogdian armour, which 
was elaborate and heavy and protected the warrior's whole body, showed advanced crafts- 

The mural paintings are a valuable source of information about feast-day customs and 
rituals, banquets, wrestling, dances and ritual bathing (Fig. 22). In connection with the 
harvest festival, an artist in Panjikent painted grain being conveyed from the threshing 
floor and a tutelary spirit of agriculture. 

Religious life 

Our knowledge of religions in Sogdiana comes from works of art, funerary monuments 
and writings - mainly Buddhist, Manichaean and Christian (Nestorian) discovered in East 
Turkestan. 21 A Christian text and Buddhist inscriptions on pottery have also been found in 
Panjikent. Buddhism, which came to Sughd from the south at an early period, flourished 
according to the Sui shu and the Tang shu. 22 By the seventh century, however, it had 
almost disappeared from Sogdiana. In the eighth century T'ang Buddhism spread among 

21 Livshits, 1981, pp. 350-62. 

22 Litvinsky, 1968. 



ISBN 978-92-3-103211-0 


Religious life 

FIG. 10. Panjikent. Mural painting. Clay on plaster. (Photo: © Vladimir Terebenin.) 

FIG. 11. Panjikent. Mural painting in location VI. Clay on plaster. (Photo: © Vladimir Terebenin.) 



ISBN 978-92-3-103211-0 


Religious life 

FIG. 12. Panjikent. Mural painting in location VI. Clay on plaster. (Photo: © Vladimir Terebenin.) 

FIG. 13. Panjikent. Statue of a dancer. Wood. (Photo: © Vladimir Terebenin.) 



ISBN 978-92-3-103211-0 


Religious life 

FIG. 14. Panjikent. Carved wood. (Photo: © Vladimir Terebenin.) 

FIG. 15. Varakhsha. General view of a painted panel. (Photo: © Vladimir Terebenin.) 

Sogdian emigrants, as a result of which most Sogdian Buddhist works are translations from 
Chinese. An inscription in the Afrasiab murals shows that in the seventh century a Sogdian 
king received assurances from foreign envoys that they were acquainted with the local 
religion of Samarkand. 23 This is clearly linked to aspirations to cultural self-determination 
during the heyday of Sughd. 

No correctly painted Buddhist images exist in Sogdian painting, but images of Hindu 
gods (of secondary importance from the Buddhist point of view) helped the Sogdians to 
create their own religious iconography in the sixth to the eighth century. As in 

23 AT 

Al'baum, 1975, pp. 52-6, translation and comments by V. A. Livshits. 



ISBN 978-92-3-103211-0 


Religious life 

FIG. 16. Afrasiab. Mural painting. (Photo: © Vladimir Terebenin.) 

FIG. 17. Afrasiab. Mural painting. (Photo: © Vladimir Terebenin.) 

Sogdian Buddhist and Manichaean texts, 24 Zurvan is depicted in the form of Brahma, 
Adbag (Ohrmazd) in that of Indra (Sakra) and Veshparkar (Vayu) in that of Shiva 
(Mahadeva). A four-armed Nana mounted on a lion (Fig. 23), a divine couple with sym- 
bols in the form of a camel and a mountain ram and other images of divinities are also 
known. The absence of highly developed forms of state organization explains the important 
role played by the worship of the divine patrons of individual families and communities. 

24 Humbach, 1975. 



ISBN 978-92-3-103211-0 


Religious life 

FIG. 18. Sughd. Silver jug (end of seventh century). (Photo: © Vladimir Terebenin.) 



ISBN 978-92-3-10321 1-0 Contents Scripts, epics and literary sources 



FIG. 19. Sughd. Silver dish (second half of eighth century). (Photo: © Vladimir Terebenin.) 

Although there were non-Zoroastrian divinities among these gods, the influence of Zoroas- 
trianism was indubitable. 25 The Sogdians probably regarded themselves as Zoroastrians, 
as indeed they were considered by al-Biruni and other authors writing in Arabic. Those 
Sogdian customs that seem contrary to Zoroastrian doctrine (Hindu-style iconography, the 
mourning of the dead) also existed in Khwarizm, whose Zoroastrianism is not open to 
doubt and where the Avestan gahanbdrs (phases of creation) were celebrated as religious 
feasts. There is evidence from the fifth century onwards in Sughd of the custom of cleaning 
the flesh from bones and burying them in ossuaries, as in Khwarizm (Fig. 24). 

Scripts, epics and literary sources 

Sogdians used different types of script according to the religion to which they belonged. 
The Buddhists used a national script of Aramaic origin, with heterograms. This script is 
also known from secular writings and from what is probably the only Zoroastrian text 
in it. 26 The Manichaeans had their own alphabet and the Christians used Syriac script, 
but both sometimes wrote in the national Sogdian script. There was also what could be 

25 Henning, 1965, p. 250. 

26 Sims- Williams, 1976, pp. 46-8; Gershevich, 1976; Livshits, 1981, p. 354. 



ISBN 978-92-3-103211-0 


Scripts, epics and literary sources 

FIG. 20. Sughd. Silver dish (second half of seventh century). (Photo: © Vladimir Terebenin.) 

FIG. 21. Mount Mug. Shield with a warrior horseman (end of seventh century). Painted wood. 
(Photo: © Vladimir Terebenin.) 



ISBN 978-92-3-103211-0 


Scripts, epics and literary sources 

FIG. 22. Panjikent. Mural painting showing ritual bathing. Location VI. (Photo: © Vladimir 

described as scientific literature in Sogdian, in particular a book about minerals, documents 
on medicine and on the calendar, and glossaries. 

A fragment of the epic of Rustam, probably translated from Middle Persian, has been 
found near Dunhuang. 27 Among the Manichaean writings, tales and fables, including some 
from the Indian Panchatantra and the Greek fables of Aesop, have been discovered. There 
are also non-Manic haean fairy-tales. The paintings of Panjikent show a similar but wider 

FIG. 23. Panjikent. Four-armed Nana mounted on a lion (seventh century). Drawing by B. I. Marshak. 


Gershevich, 1969, p. 227. 



ISBN 978-92-3-103211-0 


Scripts, epics and literary sources 

FIG. 24. Sughd. Ossuary (seventh century). (Photo: © Vladimir Terebenin.) 

repertoire of subjects from both translated and local literature: among the epic narratives, 
in addition to the story of Zohak, the tales of Rustam 28 and perhaps the Mahabharata, 


Belenitskiy, 1980, pp. 103-5, 199-202, Abb. 25-33. 



ISBN 978-92-3-103211-0 


Scripts, epics and literary sources 

there are also illustrations of several episodes of a previously unknown Sogdian epic. In 
one of the halls the pictures are accompanied by fragments of text. 29 The murals illustrate 
tales about the man who promised his daughter to a seaspirit; a prince, a bear, a wolf and 
a jackal; a wise judge (Fig. 25) and a woman's wiles; fables about a dog barking at an 
elephant, and about a blacksmith and a monkey; Aesop's fables about the goose that laid 
the golden eggs, and about the father and his sons; parables from the Panchatantra about 
the jackal, the lion and the bull; the lion and the hare; the learned men who resuscitated a 
tiger; and the foresight of the king of the monkeys. In Iran - the home of the Parthian and 
Middle Persian authors of the Manichaean works translated by the Sogdians - literary and 
folkloric parables and tales similar to those known in Sughd and recorded by the artists of 
Panjikent were popular. Similar subjects can be found in many countries. 

Direct and remarkably vivid evidence of the past is provided by the ' Ancient Sogdian 
Letters' from Dunhuang (probably written at the beginning of the fourth century) and the 
documents from Mount Mug on the upper reaches of the Zerafshan (Figs. 26 and 27). 
The Ancient Letters' describe the life of Sogdian settlers in China, while the Mug papers 
show Sughd at the time of the Arab conquest. These letters were found with legal and 
economic documents in a castle that served as the last refuge of Divashtich, the ruler of 
Panjikent, who was captured by the Arabs in 722. 30 Syriac, Bactrian, Indian (Brahmi) and 

9 ' -v ^^ ^v vr ^BV W^BWBV^BV WJBf W^b VQrf in 





fr a 


FIG. 25. Panjikent. Wall painting showing a judgment scene. Location VI. (Photo: © Vladimir 

29 Belenitskiy, 1980, pp. 116-18; Azarpay, 1981, Fig. 60. 

30 Sogdiyskie dokumenty s gory Mug, 1962-63. 



ISBN 978-92-3-103211-0 


Scripts, epics and literary sources 



i — LiT'ta _ ti+'tjir^J 

, jcnf 4Kb *J *>4 *xm 

P** ; 

r JM*' 


^*-_ — ^-t^ ?^v> £?4* -m^***— >**•** ^ 

I ^J 54A4 ^MC*H*#4^ 




U * 

" J >**^4tJ W^**<<*3 


***4-^*^t-»g ^j^jii^V^"****^ 



^^M-p^^i^ UL****** 1 * 4 

'4V-^***iV *6****S 

— ****** 




»*^*»^3SH-**«***J> iiP * ! ^* 




^O*!-^***^ 44 


^fc^H** *£****—** ***CP 




?-*<+'*'*>*-** I 

FIG. 26. Mount Mug. Sogdian document. Marriage contract. Leather. (Photo: © Vladimir Terebenin.) 



ISBN 978-92-3-103211-0 


Scripts, epics and literary sources 

FIG. 27. Mount Mug. Inscription affirming receipt of helmets and armour. Wood. (Photo: © Vladimir 

Arabic texts have been discovered in Sughd; and in Panjikent there is a Middle Persian wall 
inscription whose writer obviously came from Iran. The first evidence of the penetration 
into Sughd of the New Persian (Tajik) language, which supplanted Sogdian between the 
ninth and the eleventh century, dates to the eighth century. 



ISBN 978-92-3-103211-0 Contents Ustrushana 

Part Two 


(A/. N. Negmatov) 


Ustrushana was closely linked to Sughd ( Sogdiana) by its historical destiny and ethnic, 
linguistic and cultural history. It originally formed part of the territory of Sughd, but then 
developed its own historical and cultural identity as the area became more urbanized. Its 
rich agricultural and mineral resources, and its situation on the main trans-Asian route from 
the Near and Middle East to the heart of Central Asia, played a considerable role in this 

Ustrushana occupied a large area of the valley and steppes on the left bank of the 
middle reaches of the Syr Darya (Jaxartes), the foothills and gorges of the western part 
of Turkestan's range, the headwaters of the Zerafshan river and its principal tributaries, 
the Matcha and Fan Darya. To the west and southwest it bordered on Sughd, to the east 
and north-east on Khujand and Ferghana and to the north on Ilak and Chach (present- 
day Tashkent). Ustrushana is mentioned in the Wei shu and the Sui shu and frequently 
appears in the history of the T'ang dynasty. In the T'ang shu the region is called East- 
ern Ch'ao (the ideogram for Cao without the sign for 'water'), and Su-du-li-she-na (in 
Hsiiantsang, 629-630), and Cao (in Huei-ch'ao, 728). Early Arabic and Persian histor- 
ical sources give this name variously as Ashrushana, Asrushana, Ustrushana, Usrushna, 
Surushana and Sutrushana. The discovery and deciphering of the Sogdian documents from 
Mount Mug on the upper Zerafshan established that the correct form of the name is 
Ustrushana. 31 

By the Early Middle Ages, new towns and settlements with the characteristics of the 
rising feudal system had replaced those of the ancient period. The old capital Kurukada 
(Ura-tyube) was replaced by the city of Bunjikat (Fig. 28), 20 km to the south of the 

31 Sogdiyskie dokumenty s gory Mug, 1962, pp. 77-87. 



ISBN 978-92-3-103211-0 



FIG. 28. Bunjikat. Kala-i Kahkaha I. Plan of the city with the reconstruction of the city-walls. 

modern town of Shahristan. According to archaeological evidence, the intensive growth of 
this city began in the seventh and eighth centuries. 32 A new historical map of Ustrushana 33 
came into being, divided into a number of rustaks (regions), both on the plains and in the 
mountains, with towns such as Vagkat, Mink (in the valley of Dahkat), Shaukat (Nau), 
Kurkat, Havast, Savat and Zaamin (all of which still exist) and country settlements centred 
on castles and estates (Ak-tepe, Dungcha-tepe, Tashtemir-tepe, etc.). In the suburbs around 
the capital Bunjikat there were noblemen's castles with strong fortifications and elaborate 
architectural layouts, such as Chilhujra and Urta-kurgan. 34 

The sources give little information about the political history of Ustrushana during this 
period. The break-up of the great Central Asian states of late antiquity led to the secession 
of Ustrushana from the Sogdian federation, as recorded in the Pei-shih. From the late fifth 
to the seventh century, Ustrushana formed part of the Hephthalite and Western Turk states, 
although it probably preserved its internal autonomy and was ruled by its own kings, the 
afshins of the Kavus dynasty, some of whose names are known from written sources and 

32 Negmatov and Khmel'nitskiy, 1966. 

33 Negmatov, 1957, pp. 34^19. 

34 Negmatov et al., 1973; Pulatov, 1975. 



ISBN 978-92-3-103211-0 Contents Ustrushana 

coins. In the late seventh to the eighth century, Ustrushana was drawn into a long and 
dramatic struggle against the forces of the Arab cAbbasid caliphate. 35 


The population of Ustrushana consisted of tribes and clans, speaking a dialect of Sogdian, 
with an economy based on settled agriculture and urban crafts. The Chinese Buddhist pil- 
grims Hsiian-tsang and Huei-ch'ao note a certain community of culture - language, mores 
and customs - between the people of Ustrushana and the Sogdians of the Zerafshan val- 
ley, Ferghana, Chach and the adjacent regions. Hsiian-tsang calls the whole land between 
Suyab and Kish by the name 'Su-le', and its population 'Sogdians'. 36 Archaeological exca- 
vations show common traits in the artefacts of the region's culture. 

Agriculture, stockbreeding and mining provided a reliable threefold economic basis for 
the development of craft production, trade and urban life in Ustrushana. 37 With consider- 
able use of artificial irrigation, its people grew a wide range of agricultural produce (cere- 
als, cotton, garden crops and grapes), and bred livestock and riding and draught animals 
on the rich upland pastures. Mining output was considerable, with iron ore in the region of 
Mink, gold, silver, ammonium chloride and vitriol in the upper Zerafshan valley, lead in the 
Ura-tyube region and a number of other minerals. There was an extensive craft production 
of metal articles (such as weapons, agricultural implements, tools and household uten- 
sils), cotton, wool, silk and leather goods. Pottery manufacture was widespread, producing 
unglazed cooking pots, tableware and storage jars (for liquids and foodstuffs). There were 
potters' quarters with kilns at Bunjikat, Vagkat and Gala-tepe (fifth-sixth century). Build- 
ing, fortification and woodwork were well developed, as were wood-carving and mural 
painting. During the sixth and seventh centuries, Ustrushana's own bronze coinage was 
current in trade. 


The architecture of Ustrushana is remarkable for its variety. From the early medieval period 
(sixth-eighth century), we find structures of varied purpose and type, each with a well- 
developed, characteristic layout - royal palaces, castles of the urban and rural aristocracy, 
barracks and temples. The palace of Kala-i Kahkaha II is a three-storey building, with 
middle and upper levels set on a stepped, beaten earth platform. The first level includes an 
entrance vestibule, a stateroom, and an aiwdn (hall) opening on to a courtyard. The second 

35 Smirnova, 1981, pp. 31-5, 324-35. 

36 Livshits and Khromov, 1981, pp. 347-9, 367. 

37 Negmatov, 1957, pp. 82-112; Bilalov, 1980. 



ISBN 978-92-3-103211-0 



FIG. 29. Kala-i Kahkaha I. Reconstruction of the palace of the afshins. Drawing by S. Mamajanova. 

FIG. 30. Kala-i Kahkaha I. Palace of the afshins. Throne hall. Drawing by V. L. Voronina. 

level consists of a corridor, rooms for servants and kitchens. On the third level there is a 
suite of three staterooms, including a throne room with wide windows opening on to the 
courtyard, decorated with murals and carved woodwork. 



ISBN 978-92-3-103211-0 



FIG. 31. Chilhujra. Castle. Drawing by S. Mamajanova. 

The layout of the palace of Kala-i Kahkaha I is very complex, but clearly planned (Fig. 
29). 38 It is also built on a high platform but has a small keep, or tower, in the middle of the 
building. Its entrance in the form of an aiwan looks out on the town's rabad (suburb). An 
axial corridor divides the palace into two unequal parts. To the west there is a two-level 
hall with a throne loggia (Fig. 30) opening out at its far end and an entrance area in front, a 
second parallel 'lesser' hall and the palace shrine. To the east are found a large living room, 
a small room for servants and a separate corridor with an 'arsenal' (a store of stones). To 
the north and south, the palace had walled courtyards with kitchen, bakery and domestic 
premises. The palace gates were in the west wall of the north courtyard. 

There are two principal types of castle in the Shahristan depression. One has an elab- 
orate individualized layout with staterooms, living quarters, shrines and domestic offices, 
and an extremely rich decor of mural paintings and woodcarvings. The building is either 
positioned on a mountain crest (as in Chilhujra) (Fig. 31) or on a high man-made platform 
(as at Urta-kurgan) (Fig. 32). The other type has a simple, 'corridor-ridge' layout, with- 
out ornamentation, and is either placed on a mountain ridge (as at Tirmizak-tepe) or in an 
arable valley with fortified courtyard walls and chicanes (as at Tashtemir-tepe). 

The dwellings of early medieval townspeople have long, wide interiors, usually divided 
by partition walls into a back room, middle room and front aiwan (the northern quarter 
of Kala-i Kahkaha I). At the south wall of Kala-i Kahkaha I there is a neighbourhood of 
small separate units, detached but built close to each other. Each has a common corridor 


Voronina and Negmatov, 1975, pp. 50-71. 



ISBN 978-92-3-103211-0 




FIG. 32. Urta-kurgan. Castle. Drawing by S. Mamajanova. 

FIG. 33. Kala-i Kahkaha I. Palace. Mural painting. A she-wolf feeding two youngsters. 

leading to two or three rooms, and its own street entrance. The dwellings in the quarter 
near the city square of Kala-i Kahkaha I have a more individualized layout, which includes 
entrance aiwans, reception rooms and rooms with rich interior decoration of benches, roofs 
supported on columns and mural paintings (Figs. 33 and 34). 

The architectural ornamentation and monumental art of Ustrushana are rich and var- 
ied. Murals of high artistic quality with floral and geometric patterns and depictions of 
secular, epic-heroic, mythological and cultural scenes were an important feature of the 
interior decoration of palaces, castles and other buildings (Figs. 35-39). Many examples of 
wood-carving and clay-moulding have survived: columns, beams, cornices, friezes, thresh- 
olds, door posts and door frames, lintels, window grilles, entrance screens, artistic clay 



ISBN 978-92-3-103211-0 



FIG. 34. Kala-i Kahkaha I. Palace. Drawing of a fragment of mural painting showing musicians. 

mouldings and patterned fired bricks (Figs. 40-45). The capital, Bunjikat, was the main 
centre for the development of architecture and the applied arts. 


Ustrushana also had a rich variety of spiritual and cultural traditions, for the most part 
purely local. According to written sources, the Ustrushanians practised the so-called 'white 
religion', in which carved wooden idols were adorned with precious stones. Idols of this 
type were kept in the palaces of Haidar (the Ustrushanian ruler) in Samarra, in Ustrushana 
itself and in Buttam, and they were also brought to Ustrushana by refugees from Khuttal. 
Many toponyms in this region included the word mug (fire- worshipper). So far archaeol- 
ogists have discovered the castle and palace shrines mentioned above and an urban idol 
temple. Other finds include wooden idols in Chilhujra (Fig. 46), a house of fire at Ak-tepe 
near Nau, a dakhma at Chorsokha-tepe near Shahristan, rock burial vaults near Kurkat with 
human remains in khums (large jars) and ossuaries, and a number of other burials in khums 
and ossuaries in various regions of Ustrushana. 

All these finds are evidence of a particular local form of Zoroastrianism that incorpo- 
rated the worship of idols and various divinities and other religious practices, and is also 



ISBN 978-92-3-103211-0 



FIG. 35. Kala-i Kahkaha I. Palace. Fragment of a mural painting representing the goddess Nana on 
a lion. (Photo: © Vladimir Terebenin.) 

reflected in the monumental art of Ustrushana. In particular, the paintings of the lesser hall 
of the palace of Kala-i Kahkaha I show a three-headed, four-armed divinity, which may be 
a specifically Ustrushanian interpretation of the Hindu Vishparkar. Also depicted is a four- 
armed goddess mounted on a lion, which has been interpreted as an image of the principal 
goddess of Ustrushana and the great warrior-mother, thus personifying the worship both 



ISBN 978-92-3-103211-0 



FIG. 36. Kala-i Kahkaha I. Palace. Fragment of a mural painting representing the goddess Nana. 
(Photo: © Vladimir Terebenin.) 

FIG. 37. Kala-i Kahkaha I. Palace. Fragment of a mural painting representing a demon. (Photo: © 
Vladimir Terebenin.) 

of the productive forces of nature and of fertility, and identified with the Kushano-Sogdian 
great goddess Nana (see above, Fig. 35). 39 

The central figure of the huge composition painted on the west wall of the same lesser 
hall is of special interest. This is a large, richly dressed male seated on a zoomorphic 


Negmatov, 1984, pp. 146-64. 



ISBN 978-92-3-103211-0 



FIG. 38. Kala-i Kahkaha I. Palace. Fragment of a mural painting representing demons. (Photo: © 
Vladimir Terebenin.) 

FIG. 39. Kala-i Kahkaha I. Palace. Drawings of the fragments of mural paintings representing 

horse's-head throne. A warrior-king in a chariot, identified with this first image, is depicted 
three times on the north and east walls of the hall. These paintings are generally considered 



ISBN 978-92-3-103211-0 



FIG. 40. Kala-i Kahkaha I. Palace. Fragment of a wooden panel 'tympanum' showing the legendary 
King Zohak (lower level). 

lib *« 

■A- '" t ■ " 

FIG. 41. Kala-i Kahkaha I. Palace. Fragment of a wooden panel 'tympanum'. 

to represent an ancestor of the ruling dynasty of Ustrushana. The male figure has no appar- 
ent divine attributes and his immediate entourage includes musicians, aristocrats seated 
under a canopy, and warriors in various situations. The worship of the ancestor of a family 
line and dynasty is known from written sources. It is also known that religious as well as 
secular power was concentrated in the hands of the afshlns of Ustrushana and that they 
were almost deified, as is clear from the formula by which they were addressed - 'To the 
lord of lords from his slave so-and-so the son of so-and-so' . 40 


Negmatov and Sokolovskiy, 1977, pp. 152-3. 



ISBN 978-92-3-103211-0 



FIG. 42. Kala-i Kahkaha I. Palace. Wooden panel 'tympanum'. Detail of the medallion. 

FIG. 43. Kala-i Kahkaha I. Palace. Fragment of a wooden frieze. 

Other important features of the spiritual life of Ustrushana were extrareligious, epic- 
mythological traditions concerning Good and Evil, Light and Dark, the struggle between 
these principles and the victory of the forces of Good and Light. They were personified in 
the well-known oriental images of Kava, Faridun, Surush and Zohak; abandoned infants 
nurtured by animals (a she-wolf feeding two babies) (see above, Fig. 33); 41 a 'celestial 
musician', the harpist Zuhra (Venus); the bird- woman, Shirin; and a number of other 
images recorded in monumental painting, wood-carving and ceramics. 


Negmatov, 1977; Negmatov and Sokolovskiy, 1975. 



ISBN 978-92-3-103211-0 



FIG. 44. Kala-i Kahkaha I. Palace. Fragment of a wooden frieze. 

FIG. 45. Kala-i Kahkaha I. Palace. Drawings of wooden friezes with floral and geometrical designs. 

Artefacts reveal the characteristics of Ustrushanian culture. First, it was very traditional, 
rich and complex. Second, it was in the forefront of Central Asian cultural traditions and 
those common to the entire Eastern world. Third, in its general features and content, it 
had a number of elements close to the related Sogdian culture seen in excavations at Pan- 
jikent and Samarkand. These fit in with the general ethnic and linguistic history of Sughd 
and Ustrushana, although Ustrushanian culture developed independently and had its own 



ISBN 978-92-3-103211-0 Contents Ferghana 

FIG. 46. Chilhujra. Head of an idol. Charred wood. 


Whereas the third to the eighth century was a time of economic, national and cultural 
upsurge in Ustrushana, Ferghana (Pa-khan-na in Chinese sources) was in a different posi- 
tion. After the fall of the state of Dawan, the trend was towards territorial disintegration 
into a series of small regions and domains that experienced markedly uneven development. 
The sources reveal Ferghana's troubled political history. Although it had its own ruling 
dynasty with the title of ikhshid, their rule was sometimes interrupted. The name Alutar 
or At- Tar, a powerful king during the first quarter of the eighth century, is known. In 726 
Ferghana had two kings, one ruling over the north and subject to the Turks, the other ruling 
over the south and subject to the Arabs. From 739 onwards, all Ferghana was ruled by the 
Turk, Arslan Tarkhan. 

Ferghana occupied the whole basin, surrounded on all sides by mountains; it was rich 
and fertile and especially abundant in fruit, the famous Ferghana horses and other livestock. 
Cotton and many types of cereal were grown, and leather goods and cotton cloth were 
made. Horses, cereals, medicine, paints, glass and other goods were exported to neigh- 
bouring countries. 



ISBN 978-92-3-103211-0 Contents Ferghana 

The capital of Ferghana was first the city of Kasan, then Akhsikat on the bank of the 
Syr Darya. 42 The other towns included Urast, Kuba, Osh and Uzgend. During the seventh 
and eighth centuries, the total area of Uzgend was 20-30 ha and it consisted of a citadel 
(kuhandiz), the town itself (shahristan) and a commercial and craft quarter (rabad). It was 
particularly important as a trading post because of its proximity to the territories of the 
Turks. Osh, which consisted of a shahristan with a kuhandiz and a rabad, was regarded as 
a large and beautiful city. Rich and well supplied with water, it had markets at the foot of 
the hill. The towns of Bamkakhush and Tamakhush were situated in the valley of Isfara. 


In the western part of Ferghana, on the bank of the Syr Darya, the city of Khujand was 
going through a period of change. From the second to the fifth century, it had remained 
within the same territorial limits as during ancient times, its central nucleus occupying an 
area of approximately 20 ha. During the sixth to the eighth century, however, Khujand 
experienced a period of rapid growth and radical changes were made to its basic layout 
and fortifications, the eastern half of the old city being transformed into a new citadel 
approximately 8 ha in area. This was done using the east wall and parts of the north walls 
of the old city as foundations for the walls of the new citadel. Only the west wall was 
entirely new, as can be seen from excavations 31 and 32, from samples taken from the 
outer surfaces of the wall and from the fragments of ancient pottery in the clay of the 
early medieval walls. Parts of the former city moat were left around the east and south 
walls of the new citadel, and at the foot of the west wall a new moat was dug. The ancient 
citadel was converted into the inner palace arc of the new citadel. This early medieval 
reconstruction transformed Khujand into a large city with three main areas - the citadel, 
the town itself, and the commercial and craft quarter equipped with a mighty system of 

Khujand is mentioned in written Arabic and Persian sources in the accounts of events 
in the second half of the seventh century and in the T'ang shu's description of events of 
the second half of the eighth century (Chapter 221). According to the Arab encyclopaedist 
Yaqut, it was incorporated at an early date into the domains of the Haytal (the Heph- 
thalites). During the 680s, it was first raided by a detachment of the forces of the caliphate 
(the invaders were routed near the town). Khujand was involved in the Sogdian campaign 
against the caliphate in 721-722, when military action took place at the gates of the com- 
mercial quarter, opposite whose strong walls the invaders' catapults were set up. 

42 For a review of the sources and literature, see Gafurov, 1972, pp. 292-3; Litvinsky, 1976. 



ISBN 978-92-3-103211-0 Contents Ferghana 

During the medieval period, the territory of Khujand had its own ruler, with the title of 
malik (king). The territory was not large: apart from the city of Khujand itself, it included 
Kand and the smaller town of Samghar. Kand, which is mentioned in the early eighth- 
century Sogdian documents from Mount Mug, subsequently came to be known as Kand-i 
Bodom (town of almonds) because of the large quantities of almonds it exported to various 
countries. According to al-Muqaddasi, a river or canal ran through the bazaars of Kand. 
Samghar was in the centre of a small agricultural oasis on the right bank of the Syr Darya 
and consisted of a citadel-castle, a town and outlying buildings. The territory of Khujand 
also included several small settlements in the cultivated areas along the Syr Darya and in 
the delta part of the Khujabakyrgan. Khujand was situated on the main trans-Asian trade 
route, Kand on its Ferghana branch and Samghar on its Chach branch. This fact, together 
with access to mineral and agricultural resources, promoted the growth of these cities' 
trade and economies and also their rise to political prominence. 43 


Archaeological research has been carried out on several dozen early medieval Ferghanian 
urban and rural settlements. Varied and significant material has been obtained: (a) from the 
ruined fortifications of Kasan (a fortified citadel with chicanes in front of the gates, angle 
towers on the irregular outline of the city walls, and a castle with a mighty curtain wall 
and six towers built on a rock platform); (b) on the architecture and Buddhist religion from 
Kuba (a temple with two halls, each with its own entrance, with colossal figures of horses 
and a bearded deity with a human skull depicted on his forehead at the entrance to the 
aiwdn, and with painted clay statues representing the Buddhist pantheon in the halls); (c) 
on the construction of the castle from Kala-i Bolo in the valley of Isfara (a high platform 
with sloping sides and vertical fortress walls with loopholes, and dwellings and domestic 
offices with sloping roofs); (d) from a number of inaccessible mountain castles in the Asht 
and Isfara regions forming the defences of their river valleys; and (e) in the rural settle- 
ments (the estate of Kairagach in the valley of the River Khujabakyrgan, which has a large 
complex of buildings and a private chapel decorated with murals and pedestals bearing 
carved alabaster idols in the form of human figures with distinctive attributes, possibly 
used in the worship of family and clan ancestors). 

Interesting material has been obtained concerning the fortified settlement of Tudai- 
kalon, which is built on a high platform, with a reception room in the centre, side chambers 
and a flat roof supported on wooden columns. Among the finds is an ivory plaque depicting 

43 Negmatov, 1956, pp. 103-9. 



ISBN 978-92-3-103211-0 Contents Chach and llak 

flying goddesses of victory (Nike- Victoria), half-turned towards each other and each hold- 
ing a wreath in her hand. 44 In all, over 600 small sites (tepes, or mounds, with platform; and 
separate tepes) representing Ferghanian settlements and castles have been recorded, most 
of them belonging to the period from the third to the eighth century. 45 During the fourth 
century the culture of Ferghana's settled agricultural population reached its finest flower- 
ing and the characteristic thin-walled, red slip ceramic ware of excellent quality spread 
throughout the region. After a short period of cultural decline, the sixth to the eighth cen- 
tury saw an upsurge in the material culture of towns and settlements on the basis of new 
socio-economic conditions. 


The ethnic history of Ferghana is quite complicated. First the K'ang-chii and Sogdian, then 
Hephthalite elements were grafted on to the ancient local Saka stock, and all these elements 
combined to form the fairly cohesive population of Ferghana with its own East Iranian 
Ferghanian language. During the sixth and seventh centuries, when Ferghana became sub- 
ject to the Turks, there was increasing infiltration by Turk elements from the east and 
north, as can be seen from a group of inscriptions in runic script from Ferghana. In palaeo- 
anthropological terms, the population now belonged to the mesocranial and brachycranial 
Europoid group, with only a small percentage of dolicranial Europoids very sparsely inter- 
spersed with Mongoloid admixtures. 46 

Chach and llak 

The lands of Chach and llak gradually emerged as historical and geographical entities 
over the first half of the first millennium, although they were often given the same name 
of Chach (Shash in Arabic sources and Shi-Luo in Chinese sources). They were situated 
on the right bank of the middle reaches of the Syr Darya in the basins of its important 
tributaries, the Parak (Chirchik) and Ahangaran (Angren), and the neighbouring mountains 
of the western part of the T'ien Shan range. The economies of Chach (based on arable 
farming in the valleys and stockbreeding in the mountains) and llak (based on mining and 
stockbreeding), together with the local urban crafts of both regions, gave them an important 
role in the overall history of Central Asia. 

44 Bernshtam, 1952, pp. 233^4; Bulatova-Levina, 1961, pp. 41-3; Davidovich, 1958; Brykina, 1971; 
Saltovskaya, 1971, pp. 12-14, 20. 

45 Gorbunova, 1977, pp. 107-20; Filanovich, 1985, pp. 311-16. 

46 Klyashtorny, 1964; Livshits, 1968; Litvinsky, 1960; 1976, pp. 49-65: these works give a complete bib- 
liography of the question. 



ISBN 978-92-3-103211-0 Contents Chach and llak 

Written sources give little information about Chach and llak in the third to the eighth 
century. After the break-up of the K'ang-chii state which was centred on this region, it 
appears that lesser domains with their own ruling dynasties sprang up. During the fifth 
century they came under the supreme power of the Hephthalite state. In 606, after the ruler 
of Chach was killed and the region was incorporated into the Western Turk Kaghanate, a 
Turk tegin (ruler) was put on the throne of Chach. Under the Hephthalites and the Turks, 
however, the local autonomy of the regions continued - their rulers bore the titles of tudun 
of Chach and dihqan of llak. In the 560s, Chach was the arena of the ruinous wars of the 
Turk kaghan and the Sasanian king Khusrau I against the Hephthalite king Gatfar. One 
episode of this war ended in the capture of Chach, the Parak (Chirchik) region and the 
bank of the Syr Darya by the Turk kaghan. In the seventh century, part of the nomadic 
Tiirgesh people settled in Chach. During the first half of the eighth century, according to 
Arab sources and a Sogdian document from the castle on Mount Mug (see above, Figs. 
26 and 27), Chach, Ferghana and Sughd repeatedly formed military alliances to defend 
their territories against the Arab incursions, especially during the invasions of the forces of 
Qutaiba b. Muslim in 711, 712, 713 and 714 and of Nasr b. Sayyar in 737-738. In 739 the 
Kharijites, led by Harith b. Suraij, found refuge in Chach when harried by the forces of the 
same Nasr b. Sayyar. 


According to Hsiian-tsang, the territory of Chach was one-third smaller than that of 
Ustrushana, but the produce of both regions was the same and their peoples shared the same 
customs. The inhabitants of Chach are included in the 'List of Nations and Tribes' known 
to the Sogdians, which was found among Sogdian Manichaean texts of the eighth-ninth 
century. The people of Chach mainly followed Zoroastrian-Mazdean teachings and prac- 
tised burials in ossuaries (astodans). Buddhist preachers came as far as the territories of 
Chach, where they erected Buddhist buildings. The epic genre was widespread in Chach; 
interestingly, Firdausi states that in the Shah-name he used epic material collected for him 
by a dihqan from Chach. Chach, like Sughd, Ustrushana and Ferghana, was famed for its 
music. Dancing girls from Samarkand, Kumed, Kish, Maimurg and Chach were in high 
repute at the Chinese Imperial Court. 

Chach and llak were situated on an important sector of the trans-Asian trade route: roads 
passed through them from the Near and Middle East via Samarkand, Jizak in Ustrushana 
and Khujand, and from Central Asia through Taraz and Isfijab. The region's economic 
prosperity owed much to the caravan trade, in which silver, lead, gold, iron and copper 



ISBN 978-92-3-103211-0 Contents Chach and llak 

ware from llak were important, and also to commercial exchanges in basic necessities with 
the nomads of the nearby steppes. 47 


Archaeological investigation of the sites of Kaunchi II (late second-early fourth century), 
Kaunchi III (second half of the fourth-eighth century) and Minguruk (second half of the 
sixth-eighth century) has shown them to be representative of the culture of their period. 
Kaunchi II and III are remarkable for their advanced fortifications, and also for the build- 
ing of monumental houses, palaces and public and religious buildings, with characteristic 
methods of building domes, vaults and arched structures. At the Ming-uruk sites, citadels, 
castles and city walls were built on artificial mounds. Residential, public and religious 
buildings came to be decorated with large murals and carved clay reliefs. Burials in ossuar- 
ies began to be practised in addition to the previous custom of internment in tumuli. A 
wide variety of iron, non-ferrous and precious metalware and a variety of coinage have 
been found. 

Many towns and large mining centres grew up in Chach and llak, with a sharp increase 
in the number of towns and their geographical spread during the Kaunchi II and III periods. 
These sites show two types of town layout: geometrical and amorphous. Towns of the first 
type were probably influenced by the ancient Central Asian urban cultures, while those 
of the second type reflect the semi-nomadic life-style of agricultural and stockbreeding 
economies. On the archaeological evidence, some 100 settlements belong to this period. 
The progress of town building was accompanied by a general development of the region's 
settled agricultural life: a change to irrigated farming can be observed, based on artificial 
irrigation and the building of protective dykes and small reservoirs. Overall, the period 
was characterized by the development of crafts and trade, the exploitation of ore and raw 
materials and the growth of commercial and monetary relations. 

Thirty-two towns dating from the Ming-uruk period are known (twoand-a-half times 
more than in the previous period), although few of them were large. Greater attention was 
paid to fortification, and citadels were equipped with round and rectangular towers and with 
covered walkways. Groups of palace, religious, residential and workshop buildings have 
been studied. Most of the citadels combine the defensive, residential and administrative 
functions of a ruler's residence; some, however, were purely defensive. In the towns, groups 
of smelters', metal workers' and potters' workshops have been found. The towns were 
densely built up, with a network of streets and market places and a water supply system. 

47 Summary of information from Istoriya Uzbekskoy SSR, 1967; Gafurov, 1972. 



ISBN 978-92-3-103211-0 Contents Chach and llak 

During this period the towns of Tunkat (the capital of llak), Ulkai-toi-tepe, Ming-uruk, 
Kanka and many others expanded. There was a chain of fortresses in the Chirchik basin. 
It was also during the fourth to the seventh century that Chach grew up as a town, with a 
citadel, a ruler's palace and a shahristan. Two hundred and twenty-five archaeologically 
identified settlements throw light on the rural environment. In the Chirchik valley more 
than 30 large canals, with water-collecting installations supplying the various branches of 
the local economy, have been recorded. 48 


The complex ethnic history of Chach and llak between the third and the eighth century is 
similar to that of Ferghana. The oldest local ethnic group consisted of Saka and K'ang- 
chii tribes from beyond the Syr Darya, who were joined by large numbers of Hephthalite 
and Sogdian settlers. By the early medieval period the basic Iranian-speaking population 
of Chach and llak had been established. They probably spoke Iranian (Saka or Sogdian) 
dialects, which have left considerable traces in local toponyms and early medieval ono- 
mastics, as recorded in medieval Arabic and Persian literature and in numismatic and other 
material. The incorporation of Chach and llak into the sphere of Turk states in the sixth 
and seventh centuries led to a marked intensification of the settling and migration of the 
Turkic- speaking population. The following centuries saw the formation in Chach and llak 
of a Tajik population (speaking the presentday archaic Brichmulla dialect of the Tajiks 
on the border of the Tashkent district and southern Kazakstan) and a local Turkic popula- 
tion, just as the same process led to the formation in Ferghana of the Ferghanian group of 
northern Tajik and Turkic dialects. 49 

Classification and descriptions from Buryakov, 1975; 1982; see also Filanovich, 1983; 1985, pp. 
49 Oranskiy, 1960, pp. 63-6, 147-8, 205-10; Livshits and Khromov, 1981, pp. 247-8; Klyashtorny, 1964. 



ISBN 978-92-3-103211-0 Contents Geography and climate 




Zhang Guang-da 


Geography and climate 282 

Peoples and languages 284 

Social life and the economy 286 

Shan-shan: administrative system 288 

Political upheavals 289 

Trade 291 

Religion 292 

Geography and climate 

Between the third and the eighth century, a number of relatively large and powerful city- 
states, situated widely apart along the border of the Taklamakan desert in the Tarim basin, 
partitioned among themselves the region's oases. The Tarim basin stretches east for nearly 
1 ,600 km from the foothills of the Pamirs to the westernmost end of the Hexi (Ho Hsi) cor- 
ridor, the narrow passage along the peninsular territory of China's Gansu province through 
which ran the ancient Silk Route. With the vast expanse of the Taklamakan (337,000 sq. 
km) in the centre, the Tarim is surrounded by magnificent mountain ranges except towards 
the east where it ends in the depression of Lake Lop Nor. Along the southern edge of the 
Tarim lies the mighty Kunlun range. Farther to the southeast of the basin the Kunlun is 

See Map 6. 



ISBN 978-92-3-103211-0 Contents Geography and climate 

continued by the Altyn-tagh, which extends eastward to the Tsaidam basin in Koko Nor 
province and forms the southern boundary of the corridor. To the north-east of the Pamirs 
lies the main range of the T'ien Shan, which follows an easterly course and separates 
Dzungaria to the north from the Tarim basin to the south. 

The configuration of these ranges constitutes the most striking feature of the Tarim. 
The isolation from oceanic influences has had far-reaching effects on this enclosed region, 
producing extreme aridity and enormous temperature variations. In Central Asia, there are 
several striking instances of fluctuations in the water level of the great lakes. This led some 
explorers to believe that the climate there had also fluctuated widely. Ellsworth Huntington, 
the American geographer and a leading investigator of desiccation in the history of Central 
Asia, put forward this theory more systematically 1 to explain the terracing of lake shores 
and similar phenomena. The number of abandoned sites seemed to provide further proof of 
progressive desiccation and fluctuating water supply under the impact of cyclical climatic 
changes. These then bore directly on the interaction of nomads and sedentary peoples in 
the impulses to migration, and the rise and fall of oasis populations. 

Such an assumption is far from confirmed by the archaeological and geographic evi- 
dence, however, and there are alternative explanations. For example, Stein's careful on-the- 
spot observations of the marshy basin levels showed that large numbers of abandoned set- 
tlements were not the victims of drought. On the contrary, many failed to survive because 
of a surplus of water which damaged the irrigation systems. Others disappeared because 
of drifting sand which buried cultivation areas on the desert fringe. Nevertheless there had 
indisputably been a process of desiccation and the level of lakes and marshes had fluctu- 
ated in response to the varying annual snowfall. Given the present state of knowledge, it is 
difficult to demonstrate that the climate in the Tarim region underwent notable changes in 
historical times. Nor can a link be made between the climatic condition of the Tarim and 
the precipitation received by high ranges. 2 

Obviously, the desert could support no society: water is life in this area. Oases around 
the Tarim, such as Kashgar, Yarkand, Karghalik, Khotan, Keriya, Niya, Aksu and Kucha, 
owe their water - and with it their existence - to the snow-fed streams from the northern 
glacis of the Kunlun and the southern slopes of the T'ien Shan. These oases, in sharp 
contrast to the surrounding desert or semi-desert waste, are highly fertile and provide sites 
for human settlements. Because of the practice of irrigation agriculture, communities of 
great antiquity grew up and important city-states were formed. 

1 Huntington, 1919, pp. 359-85. 

2 Stein, 1921, Vol. 2, pp. 664-5. 



ISBN 978-92-3-103211-0 Contents Peoples and languages 

Although the water carried by the streams usually reaches far beyond the irrigated areas, 
it does not cross the great expanse of the desert, where the streams soon dry up or become 
lost among the dunes and change into a subterranean flow. The extreme points reached 
by the rivers, however, are of archaeological interest. They constitute a type of ruined 
oasis termed 'terminal oasis' by Stein. 3 Owing to the sand, which preserves what it buries, 
ancient sites can be traced far more easily in these 'terminal oases' than in other cultivated 
sites. The most striking illustrations are the numerous sand-buried sites such as those of 
Dandan-oilig, Niya, Endere and Uzun-tati. Located on the southern rim of the Taklamakan, 
they represent typical 'terminal oases' buried by moving sand during the first to the eighth 

Peoples and languages 

Archaeological and anthropological discoveries have shown that long before the gradual 
infiltration by Turks, this region was largely inhabited by Europoids and Indo-European- 
speaking peoples. From earliest times the nomadic peoples of Central Asia have constantly 
shifted their locations, some branches extending their tribal movements to the region of the 
Sita river, the modern Tarim. 

People known to the Chinese as the Sai (archaic Chinese sek) sought a home in the west 
and south of the Tarim at an early date. Patient studies of the geographic, epigraphic and 
literary evidence have identified the Sai with the Saka known to the Achaemenid Persians, 
Greek geographers and historians, and with the Sakas in Indian texts. This is an ethnic name 
like 'Yavana', 'Pahlava', 'Tukhara' and 'Cina'. The Sakas must have come to Khotan long 
before the second century B.C. and formed the kingdom of Khotan by combining with an 
indigenous people. According to the Han shu [History of the Former Han], the Sai tribes 
split and formed several states. To the north-west of Kashgar, states such as Hsiu-hsiin 
and Chiian-du were all of the Saka race. The people of Tashkurgan in the Pamirs may 
perhaps have been a branch of the Sakas, since they could have spoken a language close 
to Khotanese and Tumshuqese. Excavations have revealed Saka cemeteries in this region. 
Another Saka group settled in Tumshuq - an important site located between Kashgar and 
Aksu - and probably formed the ruling class. Documents in Saka, written on wood or 
paper and dating from the seventh to the tenth century, have been found in Khotan and 
Tumshuq. They were written in Brahmi script and represent several dialects: of those from 
Khotan and Tumshuq, the latter represents the more archaic. 4 The Tibetans called Kashgar 

3 Stein, 1907, Vol. 1, pp. 95-6, 383, 419. 

4 Bailey, 1958, p. 134. 



ISBN 978-92-3-103211-0 Contents Peoples and languages 

'Ga-hjag' and clearly the word is the same as Kanchaki, a language still spoken in villages 
near Kashgar in the eleventh century, as reported by Mahmud al-Kashgari in his Diwdn 
lughdt al-Turk (1076). It can thus be deduced that the population of the Kashgar area also 
spoke a dialect of their own which can be classified as a Saka language. 5 

More than one of the peoples who settled on the northern rim of the Tarim basin have 
been identified as speakers of the language known as Tokharian. According to the linguistic 
classification, Tokharian belongs to the centum type, or western branch of Indo-European 
languages. It shows no affinity to its immediate neighbours, belonging to one or another 
satem type of Indo-European. Manuscript fragments recovered from Kucha, Karashahr and 
Turf an, mostly Buddhist literary texts translated from Sanskrit or Central Asian languages, 
were written in the two dialects of Tokharian. Various proposals have been advanced to 
name these two dialects. They are commonly called Tokharian A and B or designated as 
Agnean and Kuchean. The former has been identified as the eastern dialect, which spread 
throughout Karashahr (Argi or Agni in ancient times) and also the Turfan region, while the 
latter, the western dialect, spread mainly in Kucha and its environs. 

Strong evidence for the existence of a third Tokharian dialect is provided by the doc- 
uments of Lou-Ian and Niya, recovered in the east and south-east of the Tarim. It is clear 
that the language of these documents in KharosthI script was not pure Gandharl or north- 
western Prakrit. It had borrowed some terms from the aboriginal dialect that was similar to 
Tokharian. It seems likely that the Krorainic of Lou-Ian, Shan-shan and Niya district was 
a dialect closely akin to Agnean and Kuchean before the spread of Gandharl as an official 
language. 6 

From the third century onwards several oasis city-states came to dominate the Tarim 
and overshadow their weaker neighbours. The oases in the south and west were separately 
united to the kingdoms of Kashgar and Khotan. Kucha, Karashahr and Kocho were consol- 
idated into independent powers to the north while the kingdom of Lou-Ian still held sway 
in the east towards Lop Nor. According to Chinese records, the name of the ruling family 
with the title of A-mo-chih (Amaca) in Kashgar was P'ei. Khotan (Khotana in KharosthI 
script, Hvatana in Brahmi and Hvamna or Hvam in the later Khotanese texts) was known 
throughout its 1,200 years as a kingdom (Hvatana-kshTra). It was founded by the Saka royal 
lineage of Visha (Vijaya in Tibetan and Yiieh-chih in Chinese), which continued to the end 
of the kingdom in 1006. The hegemony of the Khotanese over the southern oasis states 
seems to have begun in the second half of the first century. In Kucha the ruling princes 
came from the House of Po ('white'), first mentioned for the year 91 and referred to in 787 

5 Bailey, 1982, p. 56. 

6 Burrow, 1937, pp. vii-ix. See also Tikhvinsky and Litvinsky (eds.), 1988; Litvinsky (ed.), 1992. 



ISBN 978-92-3-10321 1-0 Contents Social life and the economy 

in the Chinese sources. Agni was under the rule of a royal family called Lung (Dragon) by 
the Chinese. 

Social life and the economy 

Owing to the scant information available, our understanding of the social structure of these 
oasis city-states remains fragmentary. However, since their geographic setting and many 
other aspects were similar, the written sources and archaeological material allow us to draw 
a general picture of their socio-economic life. 

The Chinese pilgrim Hsiian-tsang has left us a personal observation of Khotanese life 
and character in the fourth decade of the seventh century: 

The country [Khotan] is about 4,000 li [1 li = 0.21 A of a mile, or 0.44 km, as given by the 
explorer Sven Hedin] in circuit, the greater part is nothing but sand and gravel, the arable 
portion of the land is very limited. It is suitable for the cultivation of cereals and produces an 
abundance of fruits. It manufactures carpets, felts of fine quality, and fine-woven light silks. 
Moreover, it produces white and dark jade. The climate is soft and agreeable, but there are 
wind storms which bring with them clouds of dust. The manners and customs show a sense 
of propriety and justice. The inhabitants are mild by nature and respectful, they love to study 
literature, and distinguish themselves by their skill and industry. The people are easy-going, 
given to enjoyments, and live contented with their lot. Music is much practised in the country, 
and men love the song and the dance. Few of them wear garments of wool and fur, most dress 
in light silks and white clothes. Their appearance is full of urbanity and their customs are well 
regulated. 7 

Further information of a similar character is given in the special Notices concerning the 
Western Regions in the official Chinese dynastic histories. For example, there is additional 
(though fragmentary) material about life in Kucha. According to the Chin shu [History of 
the Chin] (Chapter 97): 

They [the people of Kucha] have a walled city and suburbs. The walls are threefold. Within 
are Buddhist temples and stupas numbering a thousand. The people are engaged in agriculture 
and husbandry. 

The Chou shu [History of the Chou] (Chapter 50) relates: 

In the penal laws [of Kucha], a murderer is executed, and a robber has one arm and one leg 
cut off. For the military and civil administrative taxes, they measure the land in order to assess 
the levies. Those who hold no fields remit in silver coins. Marriage, funerals, customs and 
products are about the same as in Karashahr. It also produces delicate felt, deerskin rugs, 
cymbals, sal ammoniac, cosmetics, good horses, wild oxen and the like. 

7 Hsiian-tsang, 1985, Ch. 12, Notice on Khotan. 



ISBN 978-92-3-10321 1-0 Contents Social life and the economy 

These descriptions are in full agreement with other early records. According to another 
report, the city walls of Kucha were indeed triple, equal in circumference to those of 
Ch'ang-an, the capital of the T'ang dynasty, and the number of stupas and temples within 
the city amounted to 1,000. We have similar descriptions of other cities. In Khotan, for 
example, Buddhist stupas were built in front of almost every house; the total number of 
stupas and temples was estimated at 1,000 and some 10,000 monks resided in the city. 

The economy of these oasis city-states was based on irrigation agriculture combined 
with livestock breeding and crafts. Two wall paintings in Kyzyl cave no. 175 depict peas- 
ants ploughing with two oxen. 8 Other iron tools were in use besides the plough, show- 
ing that the local residents had mastered advanced agricultural techniques. As cultivation 
depended entirely on irrigation, and water constituted the mainstay of economic life, the 
maintenance and periodic distribution of canal water between the various villages and 
households were strictly regulated. This activity was the primary concern and occupation 
of the royal administration. People living along the great canals were permitted only shared 
use of water and were obliged to keep the canals clean and in good repair. The T'ang even 
had a special office, the T'ao-t'o-suo, in charge of irrigation in Kucha. 9 

Wheat, barley, millet, peas, lucerne and cotton were the chief agricultural products. 
According to the reports on the Western Regions in the Liang shu [History of the Liang] 
and the Pei-shih [History of the Northern Dynasties], rice was also planted here. Melons, 
peaches, apricots, almonds, chestnuts and jujube were extensively grown. Grapes grow 
quickly and easily in this area and many people in Kucha kept wine in their houses, some 
even as much as 1,000 hu (1 hu = approx. 13.25 litres) - this wine would keep for up to ten 
years. In regions such as Shan-shan, people's livelihood depended on raising livestock, but 
stock-breeding was also pursued in Kucha and other regions. Paintings in the Kyzyl caves 
depict horses, cows, sheep and other domestic animals. 

As mentioned above in Hsiian-tsang's report, the people were luxuryloving and given 
to enjoyment. The royal palace buildings in Kucha were splendidly decorated with gold 
and jade, shining like the dwellings of the gods. The garment and adornment of adult men 
and women conformed to the Western Barbarian style. The hair-styles so characteristic of 
the donors depicted in the Kuchaean frescoes must illustrate the coiffure referred to in the 
text of Hsiiantsang. 

8 Yan Wen-ru, 1962, Nos. 7-8, p. 45. 

9 Pelliot, 1967-88: Duldur-Aqur, Nos. 57, 80, 84, 86, 90, 98. 



ISBN 978-92-3-103211-0 Contents Shan-shan: administrative system 

Shan-shan: administrative system 

The literary evidence has been supplemented by archaeological material that provides new 
data on the social life and customs of the oasis city-states. Administrative documents on 
wood found in Lou-Ian, Shan-shan and Niya, and which cover a span of 88 (or 96) years 
from the middle of the third century to the middle of the fourth, show that the state of Shan- 
shan was a monarchy. During this period, five kings - Pepiya, Tajaka, Amgoka, Mahiri (or 
Mayiri) and Vasmana - ruled in succession. 10 In the documents every king is designated 
by a lengthy title: 'Great King, King of Kings, Greatness, Victory, Right Law Staying at 
the Truth, His Majesty [mahanuava], Great King and Son of Heaven [devaputra] ' . Later, 
in the seventeenth year of Amgoka, a new title for the king appears: jitugha, jitumgha or 
citumgha. This seems to be a transcription of the Chinese title Shih-chung, transmitted to 
Shan-shan or conferred by Chinese emperors on the kings of several Western countries. 

All official orders were issued in the name of the king; he was assisted by officials who 
attended to administrative, judicial and financial affairs at both central and local level. To 
judge by his position as described in the documents, the ogu was the highest official. Other 
high titles were kitsaitsa and gusura (both of a judicial nature), kala (prince), camkura 
(protector), chu-kuo (pillar of the state) and rdjadardga (governor of the kingdom). The 
cojhbo, although inferior in rank to the dignitaries mentioned above, was an important 
and active functionary in the local administration. On a still lower level, the sothamga 
was charged with keeping the accounts of royal property and collected wine and other 
commodities, paid as taxes in kind. The vasu and ageta were the local judicial authorities. 
Divira (scribes) and lekhaharaga or dutiyae (letter-carriers) are among the subordinate 
designations that occur most frequently in the documents. 

From the mid-third to the mid-fifth century the kingdom of Shan-shan maintained its 
control over the southern route of the Tarim, leading from Dunhuang to Khotan, and incor- 
porated the smaller kingdoms and principalities of Ch'ieh-mo (Calmadana, Cherchen), 
Hsiao Yuan and Ching-chiieh (Niya, Cad'ota). At the height of its power, Shan-shan seems 
to have been composed of a series of rajas or rdyas (districts) administrated by rajadaragas 
or rdjadareyas nominated by the king. Ching-chiieh was listed, for example, among the 
rajas, retaining its original ruler. Each raja was divided into avanas and nagaras. Avana 
might first have meant a local market, but later it took on the broader meaning of market- 
town, including the land around it. Nagara (town) was sometimes used as a synonym for 
avana. The sata or sada (100 households) constituted the basic division of the avana, with 
a sadavida or karsenava at its head. It seems likely that taxes in kind (butter, wine, cereals, 

10 Rapson and Noble, 1929. 



ISBN 978-92-3-103211-0 Contents Political upheavals 

sheep, camels, carpets, felt and many other commodities) were levied by the sata and col- 
lected by the sothamgas. The year's tax was assessed both from the kilme(m)cis and from 
the rdjya. The rdjya seems to have been the land directly owned by the king, while the 
Mines were fiefs or estates granted to the nobility. Freemen owned farm land in the state 
and had the right to buy and sell their holdings. Monasteries apparently had independent 
economies and possessed their own lands. 

Apart from wooden tablets conveying royal orders, official decisions and civil and penal 
judgements, most of the Kharosthi documents on wood and leather are contracts for mar- 
riage, purchases, sales and other domestic transactions. Contracts concerning slave-trading 
are evidence of the existence of slavery. From all these documents, it appears that the soci- 
ety of Shan-shan was composed of a nobility, officials, householders, Buddhist monks and 
dajhas or dasas (slaves). 

Political upheavals 

In 445 the kingdom of Shan-shan was definitively annexed by the Wei dynasty (386-534) 
of northern China and an administrative system similar to that of China proper was intro- 
duced. However, the Northern Wei soon lost the kingdom of Shan-shan to the Juan-juan, a 
tribal confederacy that made its appearance on the steppe around the beginning of the fifth 

The rapid political upheavals experienced by the kingdom of Shan-shan were a common 
occurrence in the history of most city-states in the Tarim basin. Owing to their strategic 
geographic position, these principalities have, since the mid-third century, been subject to 
the vicissitudes of the outside world. Contacts with the Kushans, for example, profoundly 
influenced the functioning of government in some oasis states. The incessant interventions 
of the equestrian steppe confederacies sometimes brought severe trouble. At best, the oasis 
cities maintained a precarious independence, or enjoyed local autonomy in exchange for 
the payment of tribute. 

At the beginning of the fourth century, China underwent a north-south division. After 
more than two centuries of invasions and infiltration by the northern peoples, Barbarian 
dynasties were established in its northern and north-western provinces. Over a span of 135 
years (304-439), 16 ephemeral kingdoms sprang up along the northern marches of China 
proper, ruled by 3 Chinese and 13 immigrant leaders of 5 different nationalities: 3 Hsiung- 
nu, 5 Hsien-pi, 3 Ti, 1 Chieh, and 1 Chiang. In 327 the Former Liang (313-376), one of the 
Sixteen Kingdoms founded by the Chinese, with the Hexi corridor as its principal domain, 

11 Cf. Boyer et at., 1920-29; Burrow, 1940; Litvinsky (ed.), 1992. 



ISBN 978-92-3-103211-0 Contents Political upheavals 

set up a prefecture in the Turfan area on the model of the commandery system used in 
China proper. Chinese was used here side by side with the local Hu language, and schools 
were established for Confucian doctrines. 

During the fourth and fifth centuries the Former Liang and Former Chin (351-394) 
sent a number of punitive expeditions against Shan-shan, Karashahr, Kucha and Turfan. 
The mission of Yang Hsiian, general of the Former Liang, took place in 334. Four years 
later Lit Kwang, general of the Former Chin and subsequently founder of the Later Liang 
(386-403), was commissioned at the head of a force of 70,000 to conquer Kucha. In 
441-442, with the extinction of the Northern Liang (397-439) by the Northern Wei, 
Wu Hui and his brother An Chou, scions of the former royal house, fled west and occupied 
Shan-shan and Kocho. Shortly afterwards, all or part of this region fell successively under 
the control of the Northern Wei (445), the Juan-juan (413-448), the Kao-chu (?-492) and 
the Hephthalites (484-558 or 567) (see also Chapters 6 and 13). 

Between 560 and 563 the Western Turks inflicted crushing defeats on the Juan-juan and 
the Hephthalites. As a direct consequence of these victories, the Western Turks became the 
uncontested masters of the vast stretch of Eurasian steppe. The situation changed radically 
after the T'ang conquest of Kocho in 640. The decades that followed witnessed consistent 
efforts by the T'ang court to consolidate its position in the Western Regions by implement- 
ing a series of important administrative measures. Local institutions already established 
in China proper, such as the prefecture-county and canton-neighbourhood system, were 
introduced into the Qomul, Turfan and Beshbalyk districts and the Chinese equal-land sys- 
tem was put into practice in Turfan. Through a protracted and bitter struggle against the 
Tibetans (see Chapter 15), who made incursions into this area in alliance with the Turks, 
the Tiirgesh and the Karluks, the Chinese managed to regain control over the Tarim. Fol- 
lowing the traditional suzerain-vassal pattern, the T'ang court granted autonomy to all the 
oasis city-states. The local kings, retained as rulers and governors, were left in charge of 
the local administration under the supervision of the Protector- General of An-hsi (Kucha). 
For almost 150 years Kucha, Kashgar, Khotan and Suyab (replaced by Karashahr after 
719) were called the 'Four Garrisons of the An-hsi Protectorate-General'. 

The presence of these military bases, established to maintain T'ang rule over the West- 
ern Regions, led to Chinese influence in many areas of life, as revealed by recent archae- 
ological excavations. A number of Confucian classics, copied in duplicate, such as the 
Lun-yii [Analects of Confucius] with Cheng Hsiian's commentary, recently unearthed in 
the Turfan cemetery, show how deeply Chinese language and culture penetrated this region. 
It is no surprise to find that Qoshu Han, a Tiirgesh youth from Kucha who became a great 
general at the T'ang court in the mid-eighth century, was fond of reading the Han shu 



ISBN 978-92-3-103211-0 Contents Trade 

and other Chinese historical literature (T'ang shu, Chapter 104, Biographical Notice on 
Qoshu Han). 

At about the time that the T'ang dynasty was engaged in consolidating its hegemony 
over the Tarim, a great new power arose in the west: the Arabs (known to the Chinese 
as Ta-shih), who had already penetrated Transoxania and were advancing eastward. Their 
presence in Central Asia led to a clash with the Chinese. In a battle fought in July 75 1 on 
the plains near the Talas river, Kao Hsien-chih, the T'ang deputy Protector-General, was 
defeated by the Arabs under the command of Ziyad b. Salih. This disaster marked the end 
of the T'ang advance into the Western Regions. 


In spite of numerous vicissitudes, the period from the third to the eighth century saw 
the expansion of overland commerce and cultural exchange, together with the spread of 
Buddhism. The Silk Route created favourable conditions for an unprecedented increase in 
commercial and cultural activity. The most important merchandise traded along this car- 
avan road was Chinese silk. Its market value greatly exceeded that of other goods and it 
brought a substantial income both to local rulers and to residents of the oasis states. Large 
amounts of Chinese silk were exported to the West by Sogdian merchants who, from very 
early times (and to a greater extent than the nomadic peoples), were instrumental in carry- 
ing on the Eurasian caravan trade. Specimens of Chinese and Sogdian silk products have 
been found in districts as far west as Moshchevaya Balka in the Caucasus. 12 

The technique of sericulture seems to have been introduced into Khotan very early (it 
is alluded to in the legend concerning a Chinese princess who smuggles silkworms into 
the region). 13 The domestic silk- weaving industry appears to have existed in other oasis 
states too, as the terms 'Kuchean brocade' and 'Kashgar brocade' both appear in Turfan 
documents. Persian brocade (or an imitation) has been discovered in ancient tombs of 
the Astana cemetery in Turfan that date from after the seventh century. The weave is a 
weft-patterned compound twill with a double yarn warp, 14 totally different from that of 
Chinese warp-patterned design. It appears that such Persian- style brocade was specially 
manufactured for export to China proper or even to Iran. Turfan was also famous for its 
cotton textiles and almost all the oasis states were widely known for their fine carpets, felt 
products and home- woven woollen fabrics. 

12 Yerusalimskaya, 1972. 

13 Stein, 1907, Vol. 1, pp. 229-30. 

14 HsiaNai, 1979, pp. 89-97. 



ISBN 978-92-3-103211-0 Contents Religion 

Other articles transported along the Silk Route included jade from Khotan, turquoise 
from Iran, lapis lazuli from Afghanistan, tortoiseshell and ivory from India, coral and 
pearls from the Persian Gulf and the Red Sea, gold and silverware from Sasanian Iran, 
glassware from the eastern Mediterranean and bronze mirrors, tricoloured pottery and lac- 
querware from China. Other exotica, such as embroidered robes, tapestry, armour, swords, 
harness with gorgeous decorations, and metalwork, were also traded along the Silk Route. 
Byzantine gold coins and large amounts of Persian silver species have been found in many 
sites of the Tarim. 15 


Among those who used the Silk Route were Buddhist monks from the territory of the 
Kushans and north-west India travelling to the east, and itinerant Buddhists from Kucha, 
Khotan and China proper going to the west. The introduction of Buddhism into the Tarim 
basin and China is first connected with the missionary translators from Parthia and Sogdi- 
ana, and then with the expansion of the Kushan Empire, especially during the reign of the 
great ruler Kanishka I. In the earlier period of Buddhist expansion, most Buddhist scrip- 
tures were introduced to China through the Tarim, with the addition of many local cultural 
elements. One example is the Chinese translation of Buddha as Fo and Fu-Tu. This was 
simply a transliteration into Chinese of the term 'But' that was used in the Tarim basin. 16 

The two major schools of Buddhism - Hinayana and Mahayana - coexisted in the Tarim 
region, each becoming predominant at certain periods, and it is not always possible to draw 
a sharp distinction between them. 17 The earliest Buddhist texts, written mostly in Gandhari 
Prakrit, may be dated to the latter half of the second century. The discovery by Dutreuil 
de Rhins of a manuscript of a Prakrit version of the Dharmapada shows that the estab- 
lishment of Buddhism in Khotan is of considerable antiquity. The Mahayana school had 
flourished there since the beginning of the fourth century, without arousing any apprecia- 
ble hostility from the Sarvastivadins. 18 The Hinayana school was predominant in Kucha, 
where most of the wall paintings found in rock-cut vihdras (monasteries) are connected 
with the Hinayana, though, occasionally, themes belonging to the Mahayana also appear. 19 

Khotan played an important role in the propagation of Buddhism: to zealous Chi- 
nese Buddhists of the earlier period, it was the chief seat of Buddhist studies and the 

15 Hsia Nai, 1974. 

16 Bailey, 1931, Vol. 6, Part 2, pp. 282-3; Ji Xianlin et al., 1982, p. 341. 

17 Asmussen, 1965, pp. 144. 

18 For the Dharmapada, see Brough, 1962. 

19 Gaulier et al., 1976, Vol. 1, p. 24. 



ISBN 978-92-3-103211-0 



FIG. 1. Duldur-Aqur.Wall painting. Musee Guimet, Paris. Photo: © R.M.N./© Daniel Arnaudet. 

foremost source of original Buddhist texts. According to Hsiian-tsang, a large number of 
Sanskrit manuscripts of the Buddhist canons reached Khotan from Kashmir. 20 Sutras like 
the Avamamsaka appear to have been compiled in Khotan and huge quantities of Buddhist 
manuscripts were translated from Sanskrit and Gandhari into Khotanese Saka. It should be 
noted that in the later period the extreme mysticism of the Vajrayana school is also found 
in Khotan. 

The religious influence of Kucha was very strong and the region is dotted with many 
Buddhist sites (Kyzyl, Duldur-Aqur, Kyzyl-Qagha, Kumtura, Su-bashi, Simsim, 
Acigh-Ilak and Kirish) (Figs. 1 and 2). The site of Tumshuq was also a place of great 
importance in the spread of Buddhism (Figs.) Kumarajiva (344-413), one of the four great 
translators into Chinese of Buddhist scriptures, was of Kuchean origin. A monk of great 
attainments, he acquired a mastery of Chinese thought and literary style and his translations 
reached a high degree of accuracy. 

Side by side with Buddhism, pre-Zoroastrian Old Iranian beliefs continued among the 
Khotan Saka population. Remnants of the old Indo-European religion were preserved 
among the Tokharian population, and Taoist beliefs among the local Chinese. The domi- 
nance of Buddhism was also challenged by other religions. Zoroastrianism and 


Asmussen, 1965, p. 143; Litvinsky (ed.), 1992. 



ISBN 978-92-3-103211-0 



FIG. 2. Kumtura. Terracotta figurine. Musee Guimet, Paris. Photo: © R.M.N./© Droits reserves. 

Manichaeism may have entered the Tarim region much earlier than Buddhism; their activ- 
ity manifested itself in the T'ang period. According to T'ang sources, Hsien-shen ('the 
Zoroastrian god') was worshipped in Kashgar and Khotan. The term Hu-t'ien (Barbar- 
ian god) in the Chinese Turfan documents is sometimes interpreted as meaning Ohrmazd. 
Manichaeism did not prosper until the Uighurs established their supremacy in Kocho in the 



ISBN 978-92-3-103211-0 



FIG. 3. Tumshuq. Buddhist decoration. Terracotta. Musee Guimet, Paris. Photo: © R.M.N./ © Daniel 

first half of the ninth century. Sogdian Manichaean merchants must have brought their faith 
to Turfan and other states as they travelled along the Silk Route, as shown by the discovery 
of Manichaean texts in Tokharian B and by the Manichaean Sogdian Xudstvdnift fragment. 
In 635 a Syrian missionary called Aloben (known to the Chinese as A-lo-pen) introduced 
Nestorian Christianity to Ch'ang-an and established a Nestorian Church there three years 

However, traces of Nestorianism in the Tarim region are lost in obscurity (see also Chapters 
17 and 18). 

Since the Tarim basin is located at a crossroads, its culture came to reflect a peculiar 
syncretism of various heterogeneous civilizations. This can be seen clearly in Buddhist 
art. The source of Buddhist iconography in the Tarim region is to be found in Gandhara. 
Its features include the huge statues of the Buddha in various poses, the treatment of his 



ISBN 978-92-3-103211-0 



FIG. 4. Tumshuq. Buddhist decoration. Terracotta. Musee Guimet, Paris. Photo: © R.M.N./ © Daniel 

FIG. 5. Tumshuq. Buddhist decoration. Terracotta. Musee Guimet, Paris. Photo: © R.M.N./ © Droits 




ISBN 978-92-3-103211-0 Contents Religion 

hair and the drapery of his clothes, and the representation of arhats, Bodhisattvas, the 
various Buddhist divinities, ascetics, donors and other figures. The paintings of Turfan are 
characterized by an elongation of the images, a pronounced contrast of light and shade, and 
similar devices. These characteristics provide eloquent evidence that, during the Kushan 
period, Gandharan styles were mixed with Iranian, Hellenistic and Bactrian elements and 
had a predominant influence upon Buddhist art, especially during its earlier phase. This 
can be observed in the paintings of Miran in particular. Later periods even reveal traces of 
the influence of Sogdian art. 

The ruined complexes of Buddhist monastic buildings (examples include Rawak in the 
vicinity of Khotan, Duldur-Aqur and Su-bashi in Kucha, and Shorchuk near Karashahr) 
also bear silent witness to the inspiration derived from the planning of similar structures 
in India, Gandhara and Bactria. The hewing of caves or grottoes into the rocks seems to 
have been a practice borrowed from India via the Bamiyan valley and the Termez area 
(Kara-tepe). The layout of vihdras and stupas of various types around the Tarim basin also 
points to the close affinity of architectural styles. 

Music, dance and the associated fine arts, especially the Buddhist arts, also flourished 
under the influence of both east and west. Many of the musical instruments shown on the 
wall paintings of Kucha had their origin in Central Asia, India and China. The dances 
of Kucha show strong Indian influence: the swaying of the hips, the frequent changes of 
gesture and the expression in the eyes shown in the paintings are all typical of Indian dance. 
The Hu-hsiian dance, which was very famous at the T'ang court, originated from Sogdiana, 
developed in Kucha and was later introduced into Ch'ang-an, the capital of the T'ang. 
Artists of the Tarim area - such as the famous painter Yii-ch'ih I-sen (Visha Irasanga) 
from Khotan, several of whose paintings of the Devaraja still survive - introduced new 
techniques to China proper. 

Buddhist canons, religious plays and other kinds of performances were also known in 
these oasis states. Manuscripts dating from this period of the Buddhist sutra, such as the 
Dharmapada and the Uddnavarga, and of the Buddhist play the Maitreya-samitinataka, 
both in Sanskrit and in their translation into Tokharian A, have been discovered. 21 Tibetan 
manuscripts of the Li-yul lun-bstan-pa report the performance of a Buddhist play in the 
temple of Khotan. 22 

21 Liiders, 1911; Sieg and Siegling, 1921. 

22 Emmerick, 1967, pp. 41-5; Litvinsky (ed.), 1992. 



ISBN 978-92-3-103211-0 Contents KOCHO ( KAO-CH'ANG) 


Zhang Guang-da 


The Kocho Prefecture period (327-460) 300 

The Kocho kingdom period (460-640) 301 

Recently discovered documents at Kocho 302 

Administrative System and socio-economic life under the House of Ch'ii 303 

The meeting and merging of cultures 305 

The Turfan depression, located in the eastern T'ien Shan region, has an area of 50,147 
sq. km, of which 4,050 sq. km are below sea level. Lake Ayding-kol lies at 156 m below 
sea level, the lowest point of Turfan, which is bounded on the south by Mount Choi Tagh 
and on the north by Mount Bogdo Ula. These steep, rugged mountains are covered with 
snow all the year round, and the melting snows irrigate the oases of the depression's north- 
central plain. Stone and pottery relics unearthed north of the village of Astana and west 
of the moat of the ancient city of Yar (Chiao Ho) indicate that there was already human 
activity in the area more than 4,000 years ago. Over the centuries, the reputation of these 
fertile, lush and very habitable oases grew. 

The area's favourable natural conditions and its strategic location attracted many ethnic 
groups and it became a meeting point for various cultures. The Chii-shih people were early 
settlers of the Turfan depression; like the Agni of Karashahr, they presumably spoke an 
Indo-European dialect. According to Ssu-ma Ch'ien's Shih-chi [Historical Records] and 
Pan Ku's Han shu [History of the Former Han], the Chii-shih 'lived in felt tents, moved 
from place to place in search of water and pasture, and applied themselves to agricultural 

See Map 6. 



ISBN 978-92-3-103211-0 Contents KOCHO ( KAO-CH'ANG) 

work to an appreciable extent' . In 200 B.C. Turfan was bordered on the north by the territory 
of the Hsiung-nu tribe who had '3,000 trained bowmen' (Shih-chi, Chapter 123, Notice on 
Ta-yiian, i.e. Ferghana), on the south by the oasis city-states founded by the area's original 
inhabitants and on the east by the Western (Former) Han dynasty (206 b.c.-a.d. 24), whose 
capital was on the site where Xi-an stands today. In 108 B.C. the Han authorities sent a 700- 
strong cavalry unit headed by Chao Po-nu to rout the Chu-shih. 1 However, the area of 
Turfan inhabited by the Chu-shih was still ruled by the Hsiung-nu, leading to a struggle for 
control between the Hsiung-nu and the Han. This period, which lasted from 99 to 60 B.C., 
is known in Chinese history as 'the five wars for the control of the Chii-shih' (Han shu, 
Chapter 96, Notice on the Western Regions). Finally, in 60 B.C., the Han seized the area and 
appointed Cheng Chi as the first Protector-General of the Western Regions (Hsi-yii Tu-hu). 

For the next 70 years, the Han exercised firm control over the area and their culture was 
to leave a significant mark. The area was divided into the kingdoms of 'Anterior Chii-shih' 
and 'Posterior Chii-shih'. The former, which had Yar as its capital, was inside the depres- 
sion itself. The latter was situated north of Bogdo Ula, a mountain of the eastern branch of 
the T'ien Shan range. In 48 B.C. the Han stationed a wu-chi hsiao-wei (colonel of Wu and 
Chi) 2 in the Anterior Chii-shih city of Kocho (Kao-ch'ang), 30 km south-east of Turfan 
in Xinjiang. The wu-chi hsiao-wefs main responsibilities were, first, to command the Han 
troops from the central plains of China and, second, to make the soldiers work the agricul- 
tural colonies which provided food for the Han troops garrisoned in the Western Regions 
and for the Han diplomatic envoys passing through the area. During the period of Yiian- 
shih (a.d. 1-5), the wuchi hsiao-wei Hsu P'u-yii opened the 'New Northern Route', which 
greatly shortened the journey from the Jade Gate in the Dunhuang limes to the territory of 
the Posterior Chii-shih. 3 With the opening of this direct route via Hami (Qomul), Turfan 
was destined to become even more important to the Chinese than before. 

In the first century a.d., control of Turfan constantly changed hands between the Han 
and the Hsiung-nu. From 73 onwards, and especially after 89 when General Pan Ch'ao 
of the Eastern (Later) Han (24-220) brought the Tarim basin back under Han control, 
the Chii-shih were once again under Han jurisdiction. Following the conquest of Chii- 
shih, the Eastern Han re-established, after an interval of some 60 years, the offices of 
Protector- General and wu-chi hsiao-wei. Increasing numbers of Han garrison troops were 
stationed in the area and the newly opened up territory was expanded. Pan Chao's son 
Pan Yung, who was appointed chang-shih of the Western Regions in 123, stationed his 

1 Hirth, 1917, p. 106. 

2 For the title wu-chi hsiao-wei and its supposed origin, see Chavannes, 1907, p. 154, note 2; Hulsewe and 
Loewe, 1979, p. 79, note 63. 

3 Stein, 1921, Vol. 2, pp. 705 et seq.; 1928, pp. 542, 571 et seq. 



ISBN 978-92-3-103211-0 Contents The Kocho Prefecture period (327-460) 

troops at Lukchun (T'ien Ti), an important site located in the centre of Turfan, not far 
east of Kocho. Gradually, the Han Chinese from the central plains of China and the Ho 
Hsi corridor intermingled with the Chii-shih natives. During the Wei dynasty (220-265), 
founded by the House of Ts'ao, and the Western Chin dynasty (265-316), the kingdom of 
Chii-shih was basically loyal to China thanks to the implementation of a continuous policy 
of 'control by reconciliation' through the wu-chi hsiao-wei. The so-called 'Kocho soldiers' 
of the Wei and Chin dynasties may have been a local army made up of Chiishih natives and 
immigrant Han Chinese. 

The Kocho Prefecture period (327-460) 

The fall in 316 of the Western Chin led to the fragmentation and breakdown of Chinese 
power. In 327 Chang Chun of the local dynasty, the Former Liang (313-376), occupied 
Kocho and captured the wu-chi hsiao-wei Chao Chen, who had attempted to proclaim the 
independence of Turfan. The same year, Chang Chun established a prefecture in Kocho to 
administer the counties of Kocho and Lukchun. The Turfan depression and its agricultural 
colonies became a prefecture of the Former Liang. As power over the Ho Hsi area changed 
hands, so did the administration of the prefecture. Ephemeral dynasties followed and when 
in 439 the Northern Liang (397-439) were defeated by the Northern Wei (386-534), Wu 
Hui of the Northern Liang royal family led a retreat of some 10,000 families from Dunhuang 
west to Shan-shan. In 442 Wu Hui occupied Kocho, establishing himself the following year 
as king of Liang and creating a regional state that was independent of Ho Hsi district. 

In 448 Wu Hui's younger brother An Chou took control of the city of Yar and finally 
destroyed the Anterior Chii-shih regime; the remaining Chiishih forces moved westwards 
to Agni (Karashahr). Turfan's political, economic and cultural centre thus moved from Yar 
to Kocho. It was at this point that the Chii-shih people, who had been active in the Turfan 
depression since the Han dynasty, left the stage of history. In 460 An Chou 4 was killed 
in a Juan-juan invasion. As there were over 10,000 Han Chinese families living in Kocho, 
the Juan-juan - whose power base was in Mongolia - found it difficult to exercise direct 
rule and therefore placed a Chinese puppet king, K'an Po Chou, 5 on the throne. K'an Po 
Chou's reign marked the beginning of a period during which the Turfan depression was 
governed as a kingdom. 

4 The memory of An Chou is celebrated in the Chinese inscription of 469 acquired by Griinwedel from 
the ruins of a Buddhist shrine of Idikutshahri; this inscription has been edited and discussed by Franke, 1907, 
andPelliot, 1912. 

5 For the entire fifth century, some interesting notices concerning Turfan are extracted from Chinese his- 
torical sources, especially the Pei-shih, and lucidly discussed in Franke's important papers, 1907. 



ISBN 978-92-3-103211-0 Contents The Kocho kingdom period (460-640) 

Despite its remote location to the west of the central plains, the Turfan depression was 
strongly influenced by Han Chinese culture during the Kocho Prefecture period when it 
was governed by the local Ho Hsi regime. Its political and military institutions were very 
similar to those of China proper. For example, at the time when family clans held sway 
over the central plains, rich and powerful families also ruled in Turfan. Moreover, under 
the long-term influence of the wu-chi hsiao-wei system, there was almost no difference 
in status between the military and civilian populations. Civilians and soldiers recruited 
or dispatched from the central plains joined with the native inhabitants to form Kocho's 
prefectural armies which guarded the borders and built roads and canals, enabling Turfan 
to develop and prosper. 

The Kocho kingdom period (460-640) 

In 460 the Juan-juan established the House of K'an as puppet rulers. By 485 internal con- 
flicts had undermined the Juan-juan and Kocho seized the opportunity to break away from 
their rule. However, various nomadic tribes, including the Juan-juan, the Kao-chii and the 
Hephthalites, competed for power over Kocho and the independence of the House of K'an 
was shortlived. In 491 the regime was brought down by the Kao-chii nation, a powerful 
force on the steppe. The Kao-chii king set up Chang Meng Ming, a native of Dunhuang, as 
the king of Kocho. A few years later, Chang was killed by his countrymen, who put Ma Ju 
on the throne. Ma Ju, constantly harrassed by the Kao-chii, sent an envoy to the Northern 
Wei in 497, asking for permission to move his people into the interior. This caused dis- 
satisfaction among the native inhabitants. The Kao-chii killed Ma Ju, placed Ch'ii Chia on 
the throne and made Kocho the capital. Although constantly harassed by northern nomads 
and besieged by dissenters, the precarious Kocho kingdom founded by Ch'ii Chia was to 
survive for 138 years (502-640), making it the longest lived of the local regimes. 

The Kocho kingdom ruled by the Ch'ii family measured 300 li from east to west and 500 
li from north to south (Chou shu, Chapter 50, Notice on Kaoch'ang, i.e. Kocho), with about 
8,000 households and a total population of some 30,000. Kocho was the most powerful and 
most culturally advanced city-state of the western oases. Recent research by scholars into 
Turfan documents has revealed numerous reign titles (Yiian Ho, Ch'eng Ping, Lung Hsing, 
etc.) that are not recorded in China's historical annals. These findings help to clarify a series 
of political changes during the Kocho kingdom period. On the basis of detailed scholarly 
research, it is possible to plot a rough chronology of the Kocho kingdom under the House 
of Ch'ii. 6 

6 Hou, 1984, p. 75. 



ISBN 978-92-3-10321 1-0 Contents Recently discovered documents at Kocho 

The kingdom was controlled by wealthy, influential local clans closely connected 
through intermarriage. Apart from the Ch'ii family, the most eminent clans were the Changs, 
the Fans, the Yins, the Mas, the Shihs and the Hsins. At the beginning of this period, the 
tiny kingdom of Kocho was threatened by the Kao-chii. In 552 the Turk Kaghanate became 
immensely powerful, conquering the Juan-juan and dominating the northern Gobi desert. 
The kingdom of Kocho submitted to Turk rule and several Turk princes married into the 
Kocho royal family. By the beginning of the seventh century, however, the Turk Kaghanate 
had begun to decline and the T'ieh-le tribe held suzerainty over Kocho. 

During Ch'ii Po-ya's reign, the kingdom of Kocho developed close relations with the 
Sui dynasty (581-618) and actively pursued a policy of Sinicization. This caused dissat- 
isfaction among the great families, resulting in a coup d'etat. The rebel's reign lasted 6 
years, after which the Ch'ii were restored to the throne with the aid of the Chang clan, 
who were closely linked to the Ch'ii family by generations of intermarriage. The T'ang 
dynasty (618-907) won a succession of military victories against the nomads in the north 
and west, and in 630 they eliminated the powerful Eastern Turk Kaghanate. The 7 cities 
of Qomul (I-Wu), which had belonged to the Eastern Turks, were merged with the T'ang 
nation. Because his kingdom was adjacent to the Western Regions, the king of Kocho, 
Ch'ii Wen-t'ai, was fearful of T'ang ambitions to move westward and he reached an agree- 
ment with the Western Turks to resist the T'ang. When attacked by the T'ang, however, 
the Western Turks were the first to surrender; in 640 the Chii rulers of the Kocho kingdom 
also yielded. Having lasted for over 100 years and 10 reigns, the Ch'ii family's rule over 
Kocho now came to an end. The T'ang government established the Hsi-chou district and 
the An-hsi (An-xi) military Protectorate-General at Yar (Chiao Ho), 7 and for the next 152 
years (640-792) Turfan was ruled directly by the T'ang. 

Recently discovered documents at Kocho 

Following the 1898 Turfan expedition by the Russian archaeologist D. Klementz, a succes- 
sion of archaeological teams from Britain, Germany, Japan and Sweden came to this area 
to conduct excavations (Fig. 1). As the reports of these excavations were published long 
ago and are well known, we shall not dwell on them here. 8 Since 1959 other excavations 
have been conducted by Chinese archaeologists at the cemeteries of Astana and Karakhoja, 
north of the ancient city of Kocho. According to reports published between 1959 and 1975, 
13 major excavations were conducted and 459 ancient graves were opened, including 354 

7 Zhang, 1988, pp. 81-7. 

8 Dabbs, 1963. 



ISBN 978-92-3-103211-0 


Administrative System 

at the Astana cemetery. Documents were found in 1 19 of the excavated graves. Of the 1,586 
documents written in Chinese, 9 403 dated from the period when the Ch'ii family ruled the 
Kocho kingdom and 1,020 from the T'ang period. Many of the 300 epitaphs excavated 
belong to the Ch'ii period. Because of Turfan's extremely dry climate, the documents are 
well preserved and the manuscripts appear quite new. Taken together, these documents 
provide a comprehensive record of contemporary social, political, economic and cultural 
life in the Turf an region. 

FIG. 1. Kocho. Ruins of stupa 'Y'. Photo: © R.M.N./© Jean-Gilles Berizzi. 

Administrative system and socio-economic life under 
the House of Ch'u 

The Turfan documents provide a general picture of the capital, prefectures and territory of 
the Kocho kingdom, which probably derived its name from that of its capital, Kocho. The 
city was divided into at least four districts, and there were gates on all four sides of the 
surrounding city wall. The gates bore the same names as the gates of Lo-yang, the capital 
of the Han, Wei, Chin and Northern Wei dynasties; and of Gu Tsang (Wu-wei county), 
a city of the Sixteen Kingdoms. The names clearly reveal the cultural influence of China 

Tang, 1982. 



ISBN 978-92-3-103211-0 Contents Administrative System 

The Kocho kingdom under the House of Ch'ii established prefectures and counties 
modelled on the institutions that had existed in China proper during the Kocho Prefec- 
ture period. As a sop to those clans that had emigrated from China proper and from the Ho 
Hsi corridor, additional prefectures and counties were established. It is believed that the 
administrative system of this period was roughly based on a division into 4 prefectures and 
22 counties and cities, with Kocho at their centre. The most important prefectures included 
Kheng-chieh Prefecture, the political and economic centre of north-eastern Kocho, and Yar 
Prefecture, the political and cultural centre of western Kocho and formerly the capital city 
of the Anterior Chii-shih kingdom. 

During this period, the oases of the Turfan depression were fully exploited and great 
emphasis was placed on irrigation. The Kocho king had direct control over all lands, and 
every transaction involving land or vineyards required his approval. The king's permission 
was also needed when a lay person donated lands or vineyards to a temple (according to 
the documents: 'let the lay service turn into the divine one'). The king's permission was 
required even for bequests of land from father to son. These practices indicate that the 
system of land distribution in accordance with the number of people in a household - as 
implemented in the central plains of China proper - was not applied in the Kocho kingdom 
during the rule of the House of Ch'ii. 

In early times, the system of calculating wealth was to assess a family's assets by con- 
verting the household's land acreage into the amount of grain it could produce. The taxes, 
corvees and horses due from each household were determined in accordance with this 
amount. During the period of the Kocho kingdom, the system of levying taxes and corvees 
drew a distinction between monks and lay citizens, the rates probably being more moderate 
for monks than for laymen. Taxes, payable in silver coins, were levied not only in accor- 
dance with acreage, but also in accordance with the fertility of the land. The same system 
was applied to determine the quota for each of the various types of corvee. A multitude 
of other rents, corvees and trade taxes were introduced over the years. Trade flourished in 
Kocho and silver coins were used as currency. Tenancy was common during the rule of the 
House of Ch'ii. Discoveries of land lease deeds and contracts show that land was rented by 
government bodies, temples and individuals. 

The documents and epitaphs also reveal that a few influential families enjoyed a 
monopoly of power in Turfan. For example, 68 of the 300 epitaphs discovered belong to 
the powerful Chang family. Three successive generations of Changs were important offi- 
cials at court; the Changs were linked with the Ch'ii clan through intermarriage and were 
powerful enough to help the Ch'ii family regain the throne. 



ISBN 978-92-3-10321 1-0 Contents The meeting and merging of cultures 

The meeting and merging of cultures 

In Turfan, local civilizations met and intermingled with those of China, India, Iran and the 
eastern Mediterranean. Because its very foundations were the Han Chinese garrison troops 
and their agricultural colonies under the wu-chi hsiao-wei system, the Kocho Prefecture 
was profoundly influenced by Han culture. Later, when the regional governments of the 
Ho Hsi (Hexi) corridor established their prefectures in Turfan, the Han and Wei cultures, 
mingled with local elements specific to the Ho Hsi area, were introduced. During the rule 
of the House of Ch'ii, the administrative system and official titles in the Kocho kingdom 
were nearly identical to those used in the Chinese government. 10 The grave patterns, sepul- 
chral inscriptions and documents found all indicate the dominant role of Han culture; 11 for 
example, the names of the deceased and their official titles appear in Han style. Fragments 
of Mao Heng and Mao Ch'ang's version of the Book of Odes, dating from the Northern 
Liang dynasty, have been excavated at the ancient cemetery in Astana, an indication that 
the Confucian classics were in circulation there at an early time. 

Han culture rose to even greater prominence during the period when the Kocho kingdom 
was ruled by the House of Ch'ii. Ch'ii Chia, the first Ch'ii king of Kocho, petitioned the 
Northern Wei for a loan of the Confucian Five Classics and the Book of History and invited 
a Chinese man of letters (Pei-shih, Chapter 97, Notice on Kao-ch'ang) to give instruction 
to students in Kocho. Ch'ii Jian, the third king of Kocho, had a picture of 'Ai Gong, the 
duke of Lu, asking advice from Confucius' (Sui shu, Chapter 83, Notice on Kao-ch'ang) 
painted in his home to illustrate his own policy of benevolence. The books recovered from 
the Astana tombs include not only Confucian classics, but also books on history, as well 
as some poems and children's readers in Chinese such as the Ch'ien-tzu-wen [Book of 
One Thousand Characters] and the Chi-chiu-chang [Elementary Book of Chinese Charac- 
ters]. Under the Ch'ii, the ruling class actively encouraged the study of Confucian thought, 
with the aim of consolidating their own power. Confucianism thus became the dominant 
ideology of the ruling class. 

The kingdom's official records were written in Chinese characters and were similar in 
form and wording to those used in China proper during the Han and Wei periods. The 
kingdom nevertheless had its own culture. The native inhabitants wrote in the so-called 
'Barbarian script', which, judging from the existing evidence, was probably Tokharian A, 
although further study would be needed to confirm this. All in all, the influence of Chinese 
culture in Kocho was, at that time, far more profound and far-reaching than the influence 
of Buddhism from the west. 

10 Hou, 1984, pp. 52-74. 

11 Stein, 1928, Vol. 2, Ch. 19, p. 668. 



ISBN 978-92-3-10321 1-0 Contents The meeting and merging of cultures 

Taoism and metaphysics, which enjoyed great popularity in China during the Wei, Chin, 
Southern and Northern dynasties, also found their way to Kocho. Taoist talismans (vermil- 
ion characters written with cinnabar on yellow paper) have been unearthed from tomb no. 
303 at the Astana cemetery, proving that the grave's occupant was a Taoist. 12 

Because of its location on the main artery of the Silk Route, Turfan was inevitably 
influenced by Buddhism from the Kushan Empire and Tokharistan, and many Buddhist 
scriptures in Sanskrit reached Kocho from these areas. In the early twentieth century, a 
large number of hand-copied Buddhist scriptures were found in the ancient cities of Kocho, 
Yarkhoto and Sengim-aghiz by Prussian and other foreign expeditions. In the early days 
of the Kocho Prefecture period, local Buddhist monks supervised the translation into Chi- 
nese of Sanskrit Buddhist scriptures. In an ancient catalogue concerning the translation 
of the Buddhist Tripitaka (Ch'u San Tsang Chi Chi), several references are made to the 
translation of Sanskrit Buddhist scriptures by Buddhist monks from Kocho. 

Under the Ch'u, Buddhism developed rapidly. The first few kings, including Ch'u Chia 
and Ch'u Chian, considered themselves to be Confucians. Later monarchs, including the 
seventh king, Ch'u Ch'ien-ku, and the ninth, Ch'u Wen-t'ai, were pious Buddhists. Ch'u 
Wen-t'ai was especially noted for his devoutness. When on his pilgrimage to India to find 
the Buddhist scriptures, the seventh-century Chinese monk Hsiian-tsang stopped at Kocho 
and gave lectures on Buddhism. Ch'u Wen-t'ai knelt on the ground so that Hsiian-tsang 
could use his back as a step to mount the dais. 13 

In addition to the translation of Buddhist scriptures, religious activities included the 
building of temples and the carving of grottoes. The building of rock temples reached its 
peak during the period of the Sixteen Kingdoms (304- 439). To this period also should 
be assigned the carving of the Thousand-Buddha grottoes in Toyoq and Bezeklyk (Fig. 2). 
Steles erected in commemoration of the completion of temples and statues bear witness to 
the religious fervour of the imperial families. Over 40 temples were named after prominent 
families, reflecting the growing power of the family clans. In a sense, Buddhist temples 
became family shrines, indicating a process by which Buddhist power was merged with 
that of the aristocratic families of Kocho. 

The Pei-shih provides a description of the clothes worn in Kocho: 'the adult men attire 
themselves in conformity with Hu [Barbarian] style, women are in jackets and skirts and 
wear their hair in buns' (Pei-shih, Chapter 97, Notice on Kao-ch'ang). 'The adult men 
allowed their queues to fall to the back and wore long robes with narrow sleeves' (Liang- 
shu, Chapter 54, Notice on Kaoch'ang). Recent excavations at the Astana cemetery have 

12 Texts of the Documents Found in Turfan, 1981-87, Vol. 2, p. 33. 

13 Hui, 1959, Ch. 1; see also Litvinsky (ed.), 1992 and Ch. 18. 



ISBN 978-92-3-103211-0 


The meeting and merging of cultures 

& • # I 

iJ k 

FIG. 2. Bezeklyk. Buddhist grottoes. Photo: © R.M.N./© Droits reserves. 

yielded women wearing right-buttoned silk jackets and skirts and men with queues coiled 
under their necks. Many of the faces are covered with face guards (face cloths) and the eyes 
are covered by metal eye shades called 'eye cages' in Turfan documents. 14 Similar objects 
have been discovered in tombs on the Eurasian steppes as far away as Hungary, indicating 
that the use of eye shades may have been a part of the burial customs of the steppe nomads. 

Many of those with the surnames K'ang and Ts'ao had come from Samarkand, Bukhara 
and Gubdan of Zerafshan district. In Chinese historical literature, these surnames are gen- 
erally applied to the 'Nine Hu' (Barbarians) of the Chao-wu in Sogdiana district. Their 
skill as traders made them well known on the Eurasian plains throughout the Middle Ages. 
As the intermediaries of inland trade between Europe and Asia, the Sogdians also played 
an important role in the dissemination of various religions. They were actively engaged 
in the translation of Buddhist texts and played a key role in the spread of Zoroastrian- 
ism and Manichaeism. Zoroastrianism was brought to Turfan during the Kocho Prefecture 
period and the god Hu T'ien of the local pantheon may, in fact, have been Ohrmazd. The 
title sabu which occurs in documents may well have indicated an administrator in charge of 

14 Stein, 1928, Vol. 2, Ch. 19, p. 670; Texts of the Documents Found in Turfan, 1981-87, Vol. 2, pp. 61 et 
seq. See also Litvinsky (ed.), 1992. 



ISBN 978-92-3-103211-0 


The meeting and merging of cultures 

FIG. 3. Kocho. Portrait of Mani engraved on a silver dish. 

After: S. F. Ol'denburg, Russkaya Turkestanskaya eskspeditsiya 1909-1910 goda, kratkiy pred- 

voritel'niy otchet. St Petersburg, 1914. 

Zoroastrian affairs. Sogdians were also instrumental in the spread of Manichaeism. 
Manichaean texts written in Sogdian, Middle Persian, Parthian and Old Turkic have been 
discovered in Turfan (Fig. 3), in addition to a fragment in Bactrian and two fragments in 
Tokharian B. 15 However, most Manichaean texts are posterior to the influx into the Turfan 
region of the Uighurs, following the dissolution of their empire in Mongolia in 840. 

In 840 the Uighur Empire of Mongolia was overthrown by another Turkic people, the 
Kyrgyz. Fifteen of the defeated Uighur tribes fled west to settle around Kucha, Karashahr 
and Kocho and, in c. 850, finally established themselves in the Turfan basin with Kocho as 
its capital. In their new land, the Uighurs abandoned nomadism for agriculture and adopted 
a sedentary way of life. They played a major role in the Turkicization of Chinese Turkestan. 

The old religion, Manichaeism, continued to flourish but was progressively replaced by 
Buddhism. The Uighurs introduced a new cursive script adapted from the Sogdian Letters, 
which was later adopted by the Mongols. Archaeological discoveries at the beginning of 
the twentieth century have demonstrated the high level attained by Uighur culture in the 
Turfan region: a large number of the manuscripts (written in 24 different scripts) date from 
the late ninth century. Wang Yan-te, an envoy of the Sung court, visited Kocho between 
981 and 984 and was impressed by its highly colourful and civilized way of life. 


Lin, 1987; Litvinsky (ed.), 1992. 



ISBN 978-92-3-10321 1-0 Contents The meeting and merging of cultures 

In 1209 the idi-kut Barchuk, ruler of the Uighur kingdom of Kocho, voluntarily submit- 
ted to Chinggis Khan as his vassal. As the first teachers of the Mongols, the Uighur scribes 
performed a valuable service in helping the Mongols organize their administration. 



ISBN 978-92-3-103211-0 Contents NORTHERN NOMADS 


L. R. Kyzlasov 


K'ang-chu 311 

The Huns 312 

The Hsien-pi 313 

Nomad kingdoms of northern China 315 

The Juan-juan 316 

The Turks 318 

The T'ieh-le and Kao-chu 318 

Archaeological evidence 320 

In the first five centuries a.d. tribes of various origins, who were mostly herdsmen, inhab- 
ited the wide open spaces of the Eurasian steppes between the Caspian Sea in the west and 
the Great Wall of China in the east. Many of them were semi-nomadic, others were stock- 
breeders and farmers and some were agriculturalists who kept cattle. These economic dif- 
ferences resulted from the great variations in the geographical environment, ranging from 
the arid plateaux of the steppe to oases along great rivers and lakes. 

Written sources, especially in Chinese, provide the most detailed information about the 
history of the people who lived in the steppes. Unfortunately, the tumuli in the steppes 
dating back to the first five centuries have not yet been subjected to detailed archaeological 
investigation, nor are there any local epigraphic sources. 

See Map 6. 



ISBN 978-92-3-103211-0 Contents K'ang-chii 


The most extensive and stable state in the west of this region was K'ang (the ancient 
Kangha in the Avesta or K'ang-chii in the Chinese chronicles). Some scholars believe that 
the K'ang-chii state was centred on oases situated between the upper and lower reaches of 
the River Syr Darya (Jaxartes), 1 known in ancient times as the River Kanga. During the 
early period, the power of the rulers of K'ang-chii extended to the territories of Transoxania 
and the valley of the River Zerafshan, while in the north there were vassal states, the largest 
of which was Yen-ts'ai. According to the Chinese chronicles, by the second century it had 
been renamed Alania and was dependent on K'ang-chii. 2 Alania was situated between the 
Caspian Sea and the Aral Sea. 

A military and political alliance between the Sarmatian and Alan tribes living between 
the lower reaches of the Volga and the Aral Sea was formed under the name of Yen- 
ts'ai-Alania. It consisted mainly of semi-nomadic herdsmen speaking Iranian languages. 
According to Chinese sources, their customs and costume were similar to those of the 
inhabitants of K'ang-chii and their forces were 100,000 strong. The climate of their country 
was temperate, and there were many pine trees and large areas of broom and feather-grass. 
According to sixth-century sources, the Alanian region of Yen-ts'ai was renamed Su-te or 
Su-i and the Hsiung-nu from Central Asia took possession of it (apparently in the second 
century). It is reported that large numbers of merchants from Su-te came to trade in the 
Chinese region of Lanzhou, and in 564 envoys from that land came to China bearing gifts. 

In K'ang-chii itself, which lay north-west of Ta-yiian (Ferghana), although there were 
many semi-nomadic herdsmen, most of the Iranian- speaking population were reported to 
be farmers and craftsmen. The inhabitants of the region were said to lead a settled life, have 
towns, cultivate the land and breed livestock. Originally all the territories were dependent 
on the great Hsiung-nu power. The sources mention that in the first century B.C. dissent 
among the Hsiung-nu leaders weakened their power and Chih-chih (56-36 B.C.), a rebel- 
lious shan-yu (ruler) of the Hsiung-nu, sought refuge for a short time in K'angchii and 
was killed there. K'ang-chii is still mentioned in fifth-century sources, but in the sixth cen- 
tury instead of K'ang-chii we find five principalities which, as the chronicles stress, were 
situated in the 'former territories of K'ang-chii'. 3 

1 Litvinsky, 1968, pp. 14-15; Groot, 1921, pp. 5-15. See also Zuev, 1957; Hulsewe and Loewe, 1979, pp. 

2 Hulsewe and Loewe, 1979, p. 129, No. 316; see also Maenchen-Helfen, 1944^5, p. 230; Shiratori, 
1956, p. 232. 

3 Bichurin, 1950, pp. 149-275. K'ang-chii was divided into five principalities in antiquity; see Hulsewe 
and Loewe, 1979, pp. 130-1; McGovern, 1939, p. 400; Pulleyblank, 1966, p. 28; Enoki, 1956, p. 47, No. 25. 



ISBN 978-92-3-103211-0 Contents The Huns 

The Huns 

During the middle and the second half of the second century, the Greek authors Dionysius 
and Ptolemy mention the presence of Huns on the Caspian coast among the 'Scythians 
and Caspians'. In the scholarly literature, the Huns appearing on the European horizon 
are often considered a branch of the Hsiung-nu which had migrated to the West when 
the united Hsiung-nu power disintegrated in the first century. This view has been seri- 
ously challenged and has been the subject of controversy since the eighteenth century. 
According to the Roman historian Ammianus Marcellinus, the Huns, 'this restless and 
untamed people, burning with uncontrolled passion to seize the property of others, as they 
advanced, robbing and slaughtering neighbouring peoples, came to the Alans'. The Alans 
were routed, and most of them fled from the Aral Sea region and the lower reaches of the 
Volga to the northern Caucasus. There also, however, they were subject to the Huns and 
Alanian detachments were incorporated into the Hun forces. The works of Armenian his- 
torians contain hints of a struggle between the peoples of the Caucasus and the Later Huns 
in the third and fourth centuries. 

In the 370s a mass of nomadic tribes, united by the Huns in a powerful alliance, burst 
into Europe and, in 375, attacked the Eastern Goths. 4 Ammianus Marcellinus describes the 
Hun invasion in the following terms: 'This race of untamed men, without encumbrances, 
aflame with an inhuman desire for plundering others' property, made their violent way 
amid the rapine and slaughter of the neighbouring peoples. . . '. 5 The language of the Huns 
is unknown. According to Ammianus Marcellinus, they were a new tribe about which 
'ancient works know little'. He was a bitter enemy of the Huns and extremely biased in 
his descriptions of them. Nevertheless, we can deduce from his information that the Hun 
army was well organized and presented a formidable threat. Their forces were generally 
victorious and nothing could stem their advance. 

There are grounds for presuming that the Western Huns, like the Hsiungnu of Central 
Asia, had a clear-cut military and administrative system, with subdivisions into groups of 
tens, hundreds, thousands and tens of thousands. This warrior people had hereditary rulers. 

And when deliberation is called for about weighty matters [writes Ammianus Marcellinus] , 
they all consult for a common object. . . No one in their country ever plows a field or touches 
a plow-handle. They are all without fixed abode, without hearth, or lax, or settled mode of 
life, and keep roaming from place to place, like fugitives, accompanied by the wagons in 
which they live; in wagons their wives weave for them their hideous garments, in wagons 
they cohabit with their husbands, bear children, and rear them to the age of puberty. None of 

4 Maenchen-Helfen, 1973, pp. 18-168. 

5 Rolfe, 1939, p. 387. 



ISBN 978-92-3-103211-0 Contents The Hsien-pi 

their offspring, when asked, can tell you where he comes from, since he was conceived in one 
place, born far from there, and brought up still farther away. . . They dress in linen cloth or 
in the skins of field-mice sewn together, and they wear the same clothing indoors and out. . . 
They cover their heads with round caps and protect their hairy legs with goatskins; their shoes 
are formed upon no lasts, and so prevent their walking with free step. 6 

The Western Huns had lived in conditions of constant war and mass migration. When 
they reached central Europe, however, they settled in Pannonia. 

In the early 350s peoples known as the White Huns or Chionites appeared in Central 
Asia. Several scholars believe that they spoke a language of 

the Iranian group and had come south from the Aral Sea area. These warlike semi- 
nomadic herdsmen took part in raids on Sasanian Iran and even on the northern regions 
of India. They founded several principalities in Central Asia and India. Remains of their 
original towns and settlements have been found in the lower reaches of the Syr Darya (the 
sites of the Dzhety-Asar group) (for details, see Chapters 5 and 6). 

The Hsien-pi 

When the ruler of the Northern Hsiung-nu was beaten by Chinese forces in 9 1 and fled in 
an unknown direction, a new people, 

the Hsien-pi, took the opportunity to migrate, and settled on his territories. The remaining 
Hsiung-nu clans, which numbered more than 100,000 yurts, began to call themselves Hsien- 
pi, and from that time on the Hsien-pi began to gather strength. 

According to the Chinese chronicles, the Hsien-pi originated in a land of forests and high 
mountains near the basin of the River Amur. Their language and customs are described as 
similar to those of the Wu-huan, except that before a wedding they first shaved their heads, 
then held a large assembly on the river during the last month of spring; they feasted, and 
once the feasting was over, celebrated the marriage. Wild birds and beasts not to be found 
in the Middle Kingdom of China lived in the territories of the Hsien-pi, who made bows 
out of horns. There were also sables, foxes and squirrels with soft fur, from which fur coats 
renowned for their beauty were made in the Celestial Kingdom. The breeding of cattle, 
sheep, goats and horses by the Hsien-pi is also mentioned and they are said frequently to 
rustle each other's herds of livestock and horses. 

The Hsien-pi were described by one of the Chinese emperor's councillors in 117 as 

After the Hsiung-nu fled, the Hsien-pi, who took over their former territories, grew in strength. 
They have hundreds of thousands of warriors, they are remarkable for their physical strength, 

6 Ibid., pp. 383-7. 



ISBN 978-92-3-103211-0 Contents The Hsien-pi 

and are more quick-witted than the Hsiung-nu. It should also be noted that, as a result of lack 
of discipline at the guard-posts on the line of fortifications, there are many ways of evading 
the embargo, which robbers use to obtain fine metal and iron of good quality. The Chinese get 
in [through these gaps] and become the main counsellors of the Hsien-pi, and so they acquire 
keener weapons and faster horses than the Huns. 

During the reign of the Han emperor Huang-ti (146-168), an energetic leader named T'an- 
shih-huai appeared among the Hsien-pi. He subjected the elders to his authority, introduced 
laws, gathered large forces and defeated the Northern Hsiung-nu around 155. 

All the elders of the eastern and western nomadic communities submitted to him. As a result 
of this he looted the lands along the line of fortifications, repulsed the Ting-ling in the north, 
made the Fu-yii kingdom retreat in the east, attacked the Wu-sun in the west, and took pos- 
session of all the former Hsiung-nu territories, which extended for more than 14,000 li to the 
east and the west, were intersected by mountains and rivers, and had large numbers of fresh 
and salt water lakes. 

Thus the territories of the Hsien-pi extended as far as those settled by the Wusun in the Hi 
basin in the west, while in the north they adjoined those of the Ting-ling alliance of tribes 
which occupied the Altai mountains, the basins of the upper and middle Yenisey and the 
areas adjoining and to the west of Lake Baikal. 

During the Hsien-pi period, culture in Central Asia declined and society regressed com- 
pared with the Hsiung-nu state which preceded it. Many towns and settlements which had 
flourished during Hsiung-nu rule appear to have died out, craft production declined and 
centres of agricultural activity vanished (or at least the written sources which describe the 
Hsien-pi make no mention of them). 

During their heyday, the Hsiung-nu shan-yiis had been recognized as Sons of Heaven 
and the equals of the emperors of Han China, and Chinese historians considered their state 
to be comparable in strength and power with the Middle Kingdom. When the might of the 
Hsien-pi was at its greatest, the emperor Huang-ti is reported to have sent an envoy with a 
seal and cord granting T'an-shih-huai the title of wang (prince) and seeking to conclude a 
peace treaty with him on the basis of kinship. T'an-shih-huai unhesitatingly rejected these 
advances, refusing to accept a tributary relationship with the Han emperor. 

The history of the Eastern (Later) Han dynasty (24-220) is full of information and 
complaints about the Hsien-pi raids on the frontier districts and territories of China. These 
areas suffered greatly from robbery, mass murders and the abduction of vast numbers of 
people into captivity. One of the sources gives the following explanation for these raids: 

The number of the Hsien-pi increased every day, and stock-breeding and hunting could no 
longer satisfy their needs for food. T'an-shih-huai therefore rode out to inspect his lands. He 
saw the River Wu-huo-ching, extending for several hundred li. There were large numbers of 



ISBN 978-92-3-10321 1-0 Contents Nomad kingdoms of northern China 

fish in the creeks, but the Hsien-pi did not know how to catch them. When he heard that the 
inhabitants of Vozhen were skilled in catching fish with nets, T'an-shih-huai attacked this 
land on the east, captured over 1 ,000 families and resettled them on the banks of the River 
Ukhotsin, ordering them to catch fish in order to make up the insufficiency of food. 

It is clear that the Hsien-pi did not even consider engaging in agriculture or crafts. 7 

There is no evidence that the Hsien-pi expanded to the west or north. All their efforts 
appear to have been directed to the south, to the rich districts of northern China. By the 
end of the rule of the Eastern Han in 220, the Hsienpi, together with other nomad armies, 
had advanced as far as the basin of the River Liaohe and some of their tribes (A-zha in 
Tibetan literature) had even migrated to Gansu and Chinghai. Tens of thousands of Hsien- 
pi had settled over the Central Plain and other inner regions of China by the end of the third 
century. The largest ethnic groups among them were the Mu-jung, T'o-pa (Tabgach) and 

Nomad kingdoms of northern China 

These nomadic settlers enjoyed great military strength and founded their own kingdoms in 
northern China. Of the various kingdoms established by the Hsienpi in northern China, the 
Northern Wei (386-534), founded by the T'o-pa leader T'o-pa Kui, became particularly 
strong. T'o-pa Kui is said to have pacified the people and devoted his attention to agricul- 
ture. The Hsien-pi who settled in the Northern Wei kingdom rapidly made the transition 
from a patriarchal slaveowning society to the feudal system. They later assimilated with 
the Chinese. 8 

The Late Hsiung-nu realm of Yuen-pan, described by the sources as situated 'to the 
north-west of Wu-sun' (in the present-day district of Tarbagatai), was one of the districts 
belonging to the shan-yu of the Northern Hsiung-nu. In 93, when the Northern shan-yu had 
migrated westwards over the mountains to K'ang-chii: 

the weaker nomads who were not up to following them remained in the north of Kucha. 
They occupy an area of several thousand li, and number up to 200,000. . . Their customs 
and language are the same as those of the Kao-chii [i.e. Turkic-speaking tribes], but they are 
better groomed. . . They trim their hair and make their eyebrows even, applying a paste to 
them which makes them glossy. They wash three times a day before eating. 

7 Taskin, 1984, pp. 70-80. 

8 Shan, 1959, pp. 138-41. 



ISBN 978-92-3-103211-0 Contents The Juan-juan 

The Juan-juan 

The inhabitants of Yueh-pan waged war against the Juan-juan. In 449: 

their ruler sent an envoy to the court with gifts, and he also sent a remarkable physician. . . 
It was said that in their state there were sorcerers who could produce long periods of rain, 
great storms and even flooding during Juan-juan attacks. Two-tenths - or perhaps as many as 
three-tenths - of the Juan-juan drowned or died of cold. . . Afterwards the ruler always sent 
envoys with gifts. 9 

The Juan-juan became known as a distinct ethnic group from the end of the third cen- 
tury. Constantly attacked by the Wei kingdom, the Juan-juan manoeuvred in the Gobi until, 
at the end of the fourth century, they overcame the Kao-chii (see pages 323 et seq.) who 
lived to the north of the desert. Their ruler, She-lun, settled on the River Khalkha. 

Here for the first time he established military laws, according to which 1 ,000 men formed 
a detachment under an appointed commander, and 100 men made up a 'banner' under an 
appointed leader. The prisoners and booty taken were granted to the man who first broke 
into the ranks of the enemy. . . They had no written alphabet, so that they could not keep 
written records, but later they learnt to make records well by making notches in wood. . . 
She-lun earned the epithet of powerful and prosperous. He bred livestock, moving from place 
to place, wherever he could find water and grass. 

Further to the west of his territories were the lands of Yen-ch'i [Karashahr], and the lands 
of Ch'ao-hsien; in the north his realm occupied all the sandy desert and reached Hanhai [the 
upper reaches of the Amur], and in the south approached the Great Desert. He held all the 
small countries as if they were on a leash, and they were subject to him. Because of this, 
She-lun assumed the local title of ch'iu-tou-fa kaghan: in the language of the Wei dynasty, 
ch'iu-tou-fa means 'ruling and leading to expansion' and kaghan means 'emperor'. 

Around 400 the Juan-juan established a powerful empire (402-555) in Mongolia. 10 
From 402 onwards, Juan-juan forces made regular incursions on the frontier districts of 
northern China; in the ensuing wars, which lasted for several decades, sometimes the Juan- 
juan were victorious and sometimes the Northern Wei. At the beginning of the fifth century 
the Juan-juan repeatedly attacked the Wu-sun state, situated in Semirechye, driving the 
local tribes of herdsmen out to the Pamir mountains. 

According to the Chinese chronicles, the Juan-juan: 

graze their livestock, going from place to place in search of water and grass. They live in 
dome-shaped huts. They plait their hair. They wear narrow-sleeved silk robes with woven 
patterns, tight trousers and high waterproof boots. In their land they suffer from cold, and as 
early as the seventh moon ice-floes float on the rivers, blocking their course. 

9 Bichurin, 1950, Vol. 2, pp. 258-60. 
10 Sinor, 1969, pp. 97-9. 



ISBN 978-92-3-103211-0 Contents The Juan-juan 

In their realm they use sorcery to offer sacrifices to heaven and call up a wind that brings 
snow. [As a result] , ahead of them the sun shines brightly and behind them there are streams 
of muddy water. Because of this, when they are defeated it is impossible to catch up with 

Another chronicle reports: 

They do not have towns surrounded with inner and outer walls, but herd livestock, going 
from place to place in search of water and grass. Their homes are felt tents, which they take 
to the place where they stop. There is no green grass in the steppes, the climate is cold, the 
horses and cattle chew dry grass and lick the snow, but are naturally fat and strong. The 
administration of the state is simple. There are no official written documents, and they keep 
records by making notches in wood. 

The Juan-juan kagham and nobles were well acquainted with Buddhist teachings and were 
probably Buddhists as early as the beginning of the sixth century. It is known that in 511 
they sent a Buddhist monk and preacher to China with the gift of an image of the Buddha 
ornamented with pearls for the emperor. It was at this time that the Juan-juan are reported to 
have first built a town: they surrounded it with inner and outer walls and called it Mumo- 
chen. They also gradually learnt to write, and by now there were many learned people 
among them. 

It may be assumed that by then some of the Juan-juan already lived a settled life and 
practised agriculture. The original sources repeatedly mention that their kaghans obtained 
'seed millet' from China (some 10,000 shi each time). This shows that the Juan-juan soci- 
ety and state had gradually developed from nomadic herding to a settled agricultural way of 
life, from yurts to the building of houses and monumental architecture, from the nomadic 
district to towns. They had invented their own system of writing and developed their own 
local culture and Buddhist learning flourished. 

The Juan-juan state was undoubtedly multi-ethnic, but there is no definite evidence as 
to their language. As the ancient sources regard the Juan-juan as a separate branch of the 
Hsiung-nu, it may be assumed that the Juan-juan language belonged to the same linguistic 
family as that of the Hsiung-nu (whose language is also unknown). Some scholars link 
the Central Asian Juan-juan with the Avars (see also page 323) who came to Europe from 
the east in the mid-sixth century. According to a widespread but unproven and probably 
unjustified opinion, the Avars spoke a language of the Turkic group. 

During the late years of their rule in Central Asia, the Juan-juan contracted to guard 
the frontiers of northern China, with whose court they were allied by marriage. In 538 the 
daughter of the Juan-juan kaghan A-na-kui became empress of the Western Wei kingdom 
(535-556). In 535 a princess from the Eastern Wei kingdom (534-550) married A-na-kui, 



ISBN 978-92-3-10321 1-0 Contents The T'ieh-le and Kao-chii 

and in 545 the real ruler of the Eastern Wei married another of A-na-kui's daughters. A rel- 
ative peace was thus established on the frontiers of northern China. The previous constant 
wars with China had exhausted the human and economic resources of the kaghanate and 
led to internal revolts and risings of the peoples subjugated by it. 

The Turks 

Among these peoples the Turks, who lived on the Altai and supplied the Juan-juan with fer- 
rous metal products such as iron blooms, tools and weapons, became particularly powerful 
in the 530s. The rising of the Turks, who at about that time were joined by the T'ieh- 
le nomads (with up to 50,000 wagons), was crowned with success. In 552 the Juan-juan 
kaghan A-na-kui was routed by the Turks and committed suicide. 11 In the period before 
555 the Turks and the Chinese had killed large numbers of Juan-juan who were fleeing to 
China and westward towards the Aral Sea. A new state, the Turk Kaghanate (552-630, 
683-745), was established in Mongolia (see Chapter 14). 

It was precisely at this time that the first information about the Avars being pursued by 
the Turks appears in Western chronicles. In c. 562, for example, the Byzantine historian 
Menander Protector 12 wrote that Silziboulos, the ruler of the Turks, having learnt of the 
Avars' retreat after an attack on the Turks, sent the following message to Byzantium: 

The Avars are not birds, to escape Turk swords by flying through the air; they are not fish, to 
dive into the water and disappear in the depths of the sea; they wander over the surface of the 
earth. When I finish the war with the Hephthalites I shall attack the Avars, and they will not 
escape my forces. 

According to Menander, in 568 the emperor Justin II asked a Turk who was visiting 
Constantinople: 'Tell us how many Avars have cast off Turk rule, and whether you still 
have any Avars.' He was told that: 'There are Avars who are still faithful to us; we suppose 
that up to 20,000 have fled from us.' The Turks also called the Avars 'Ouarchonites' and 
regarded them as their subjects. As mentioned previously, from these facts some scholars 
conclude that the Avars, sometimes known as Ouar, Koun or Ouarchonites, are the same 
people as the Juan-juan. 

The T'ieh-le and Kao-chu 

The T'ieh-le group was the strongest and largest of the various tribes subject to the Juan- 
juan Kaghanate. According to the Chinese chronicles, of the 15 tribes belonging to the 

11 Taskin, 1984, pp. 267-95. 

12 Blockley, 1985. 



ISBN 978-92-3-10321 1-0 Contents The T'ieh-le and Kao-chii 

T'ieh-le, some were the descendants of the Hsiung-nu. Those tribes whom the Chinese 
called Kao-chii (High Chariot) were regarded as the closest to the T'ieh-le in terms of 
ethnic composition. The Chinese sources provide a fairly detailed picture of the Kao-chii, 
who were apparently the last surviving branch of the ancient Chidi. Originally known as 
the Ting-ling, in the north they were called the Chi-lei, and in China the Kao-chii Ting-ling 
(High Chariot Ting-ling). 

The Kao-chii were constantly at war with the Juan-juan and also frequently attacked 
and plundered the borders of the Wei state. In 397 the Kaochii, together with the Juan- 
juan, became subject to T'o-pa Kui, who founded the Northern Wei dynasty. When T'o-pa 
Kui declared himself emperor the following year, the Kao-chii confirmed their subjection 
to him. At the end of the fourth century their territories situated to the north of the 'sandy 
steppe' (i.e. the Gobi) were seized by the first Juan-juan kaghan, She-lun (402-410). 

From the end of the fourth to the beginning of the fifth century the Northern Wei 
launched 9 successive campaigns against the Kao-chii, taking prisoners and reportedly 
seizing over 200,000 head of horses, cattle and sheep. Finding themselves between two 
fires, the Kao-chii were later forced to surrender to the Northern Wei forces. Several hun- 
dred thousand yurts, with more than a million head of horses, cattle and sheep, were reset- 
tled to the south of the desert. 

The Kao-chii later tried to exploit the disagreements between the Juan-juan and the 
Northern Wei kingdom to gain their independence. In 487 the Kao-chii leader A-fu-chi- 
lo raised a revolt against the Juan-juan and migrated westwards at the head of 100,000 
warriors. He established a kingdom (487-541) to the north-west of the present-day Turfan 
and there he declared himself a wang (prince). In his message to the Northern Wei court, 
A-fu-chi-lo describes the situation in these terms: 

The Juan-juan rob the Son of Heaven. I exhorted them, but they would not listen to me, and 
therefore I stirred up a revolt, travelled to the territory I now occupy, and declared myself 
ruler. In the interests of the Son of Heaven I must punish and annihilate the Juan-juan. 

The Kao-chii state lasted for 55 years and won a number of victories over the Juan-juan 
forces, but quarrels broke out among the Kao-chii nobles. A-fu-chi-lo was killed and Mieh- 
tu was made ruler. On his accession to the throne, Miehtu began to send tribute to the court 
of the Wei dynasty again. 

War then broke out with the Juan-juan, in which the latter were victorious and killed 
Mieh-tu. The Juan-juan kaghan Ch'ou-nu covered Mieh-tu's skull with black lacquer and 
made it into a drinking goblet. During the reign of the new rulers of the Kao-chii, I-fu, 
Yiieh-ch'u and Pi-shih, war between the Kao-chii and the Juan-juan continued from 520 to 
542, with success now on one side, now on the other. 



ISBN 978-92-3-103211-0 Contents Archaeological evidence 

Despite all these failures, the Kao-chii, who are known as the T'ieh-le in later Chinese 
chronicles (see the Tang shu, for example), did not give up the idea of founding their own 
state. In 536 the first Turk kaghan, Bumin, attacked the T'ieh-le and captured as many 
as 50,000 wagons. Even after the incorporation of the T'ieh-le into the Turk state, they 
continued their struggle for free-dom. They waged war with the Turks from 602 to 605, 
and in 618 were subjugated by the Turk ruler, T'ung yabghu. The chronicles state that 'the 
Turks performed feats of valour with their forces in the deserts of the north' . 13 

The T'ieh-le tribes were numerous. The best known among them at a later stage were 
the Hsieh-yen-to, the Qurigan and the Uighurs, who created their own states. Although 
they spoke Turkic languages, their origins and culture differed from all the other Turkic- 
speaking peoples of the Middle Ages. During the early sixth century the entire Eurasian 
steppe zone came under the power of the Turk Kaghanate. Turkic gradually took the place 
of several Iranian languages, some of which ceased to be used. 

Archaeological evidence 

The archaeological remains of the ethnic groups which inhabited the steppe zones in the 
first five centuries a.d. such as the Hsien-pi, Yuen-pan, Juan-juan and Kao-chii or T'ieh-le 
have not yet been studied in depth. However, remains from the first to the fifth century, 
presumably connected with the Western and White Huns, are partially known to archae- 
ologists. It appears that their nobility wore richly adorned garments in what is called the 
polychrome style, made of gold and decorated with many inset semi-precious stones and 
patterns in cloisonne enamel. Examples of richly decorated weapons, harness, 'Hun-type' 
cast bronze cauldrons and other articles have been found far to the west. 

The tombs of nobles, containing articles in the above-mentioned style, have been dis- 
covered over a wide area reaching from the T'ien Shan and the Altai in the east to Pan- 
nonia and the sources of the Danube in the west. The most thoroughly studied antiquities 
are those of Hungary, the Danube area and the steppes adjoining the Black Sea to the west 
of the Volga. As for the northern region of the steppes, we have some knowledge of the 
antiquities of the areas around the lower Volga, the Aral Sea and the lower reaches of the 
Syr Darya, where the Alans and White Huns lived. 

The wide steppes of Kazakstan, Mongolia and Dzungaria, however, are still blank 
spaces in the archaeological map of the first five centuries. Much further investigation 
of both archaeological and written sources will be required before we know where the var- 
ious northern semi-nomadic peoples lived and understand their economy, way of life and 

13 Bichurin, 1950, Vol. 1, pp. 228, 243, 279, 283, 301; Taskin, 1984, pp. 278-9, 401-6. 



ISBN 978-92-3-103211-0 Contents THE TURK EMPIRE 


D. Sinor and S. G. Klyashtorny 


THE FIRST TURK EMPIRE (553-682) 322 

Ethnogenesis 322 

The economy 326 

Political history 327 


Resurgence of the Turk Empire 330 

Political and social structure 331 

Relations with China 333 

The empire in crisis 334 

The last war with T'ang China 335 

The final decade 336 

Epigraphic memorials of the Turks 337 

The Turgesh state 341 

The Uighurs and the Karluks 342 

See Map 7. 



ISBN 978-92-3-103211-0 Contents Ethnogenesis 

Part One 


(D. Sinor) 

The two centuries during which the Turks were the dominant power in Inner Asia would 
seem to mark a turning point since, for the first time in recorded history, an essentially 
nomad empire bordered simultaneously on three major sedentary civilizations: those of 
China, Iran, and the Western world as represented by Byzantium. A more or less permanent 
link was established between these three civilizations, allowing the free flow of trade and 
with it, one must presume, a range of ideas and information. 

There are other reasons for attaching great importance to the emergence of the Turks: 
not only were they the first Altaic people to leave behind indigenous historical documents; 
they were also the first Altaic people to leave behind documents written in an Altaic lan- 
guage, namely Turkic - these constitute the earliest textual evidence of any Altaic lan- 
guage. It follows from these two points that the Turks were the first people to form a major 
nomad empire centred on present-day Mongolia whose language can be established with 
absolute certainty. The attribution of a given language to any of the earlier great nomad 
empires (such as, for instance, that of the Hsiung-nu or the Juan-juan) remains highly 
speculative. Finally, the Turks became the eponymous people of all the Turks who fol- 
lowed them throughout history. We are particularly fortunate in that the history and civi- 
lization of the Tiirks can be studied through a variety of written sources, including Chinese, 
Persian, Armenian, Greek and Latin texts, in addition to the indigenous Turk or Sogdian 


In terms of political history, the Tiirks entered the scene in 552 with the revolt of the Turk 
kaghan (chief) Bumin, who overthrew the Juan-juan Empire of which his people had, up 



ISBN 978-92-3-103211-0 Contents Ethnogenesis 

to then, formed an integral part. For the period preceding this fateful event, we must rely 
on the (often self-contradictory) testimony of Chinese sources. 

According to the Chou shu, 'No doubt the Turks are a detached branch of the Hsiung- 
nu,' an opinion taken over verbatim by the Pei-shih. But the Chou shu also relates 'another 
tradition' according to which the Turks 'originated in the country of So, located north of the 
Hsiung-nu'. Since the location of So cannot be established, the information is of little use 
and simply shows that, according to this 'other tradition', the Turks were not a part of the 
Hsiungnu confederation. Moreover, Chinese sources are wont to attribute Hsiung-nu origin 
to any people belonging to the vast group of Northern or Western Barbarians. Indeed, such 
an indication may almost be considered a simple stylistic device, just as Greek sources 
would attribute Scythian origins to any nomad people appearing on the steppe. 

More importantly, Chinese sources record at least three different legends concerning 
the origin of the Turks. The first of these, which we may call that of 'The Abandoned 
Child Brought up by a Wolf , is related with slight variations by both the Chou shu and the 
Pei-shih. It tells the story of a young boy mutilated by the enemy and thrown into a marsh 
where he has intercourse with a she-wolf. The wolf and the boy subsequently take refuge 
in a cavern, where the wolf gives birth to ten boys. Several generations later the Turks 
emerge from the cavern and become the blacksmiths of the Juan-juan. There is another 
legend, also related in the Chou shu, which, in the words of this source, 'differs from the 
other [legend], nevertheless it shows that [the Turks] descended from a wolf. 

A third legend is preserved only in a collection of anecdotes, curious and miraculous 
histories probably compiled in 860 and entitled the Yu-yang tsa-tsu. According to this 
legend, which we may call that of 'The Spirit of the Lake', the ancestor of the Turks, 
who is called She-mo-she-li and lives in a cavern, has a liaison with the daughter of the 
lake spirit. One day, as the Turks are preparing for a great hunt, the girl says to She-mo: 
'Tomorrow during the hunt a white deer with golden horns will come out from the cavern 
where your ancestors were born [author's emphasis]. If your arrow hits the deer we will 
keep in touch as long as you live, but if you miss it our relationship will end.' In the course 
of the hunt, a follower of She-mo kills the deer. She-mo angrily decapitates the culprit 
and orders that a human sacrifice be established in which a man of that follower's tribe be 
beheaded. According to the Yu-yang tsa-tsu, the sacrifice remained in practice 'to this day'. 

There is no reason to impugn the authenticity of these legendary traditions, which 
clearly reveal the composite ethnic character of the Turks. The three legends differ in so 
many essential points (which cannot be examined here in detail) that they cannot possibly 
represent a single tradition. 



ISBN 978-92-3-103211-0 Contents Ethnogenesis 

The theme of the 'wolf in two of the three Turk legends is shared with the Wu-sun, who 
preceded the Turk Empire by many centuries. Also shared with the Wu-sun is the theme of 
the mutilated child abandoned in the wilderness by the enemy. According to the Shih-chi, 
the Wu-sun ruler K'un-mo was cast out to die when still a baby, but was nourished by 
birds that brought him meat and by a wolf that suckled him. The story is also related in 
the Han shu and its close relationship with one of the Turk origin myths is obvious. There 
is, however, the significant difference that, whereas in the Wu-sun myth the wolf saves the 
ancestor of the tribe, it is not - as in the case of the Turks - the ancestor of the people. (The 
connections with Mongol myths, though undeniable, should not concern us here.) 

The theme of the 'cavern' that appears in two of the three Turk ancestral legends has 
its parallels in later Mongol mythology. More importantly, it establishes a link between 
the Turks and the Kyrgyz, whose ancestor, according to the Yu-yang tsa-tsu, 'lived in a 
cavern to the north of the Kogman mountain' . According to the same source, however, 'the 
Kyrgyz do not belong to the race of the wolf. 

There is convincing evidence to show that Turk ceremonial practices took into account 
the existence of these two themes, namely those of the cavern and the wolf. For example, 
the Pei-shih clearly states that, 'In front of the gate to the camp [the Turks] placed a stan- 
dard with a wolf's head on it, so as to show that they had not forgotten their origins.' This 
is confirmed by the Chou shu: '[The Turks] put golden wolves' head on their standards. . . 
The Turks descended from a wolf and did not want to forget their origin.' A basrelief 
on the Sogdian Bugut inscription erected by Turk rulers represents a she-wolf with a small 
human figure under her belly and thus supports the evidence of written sources. The present 
author ventures the hypothesis that the 'wolf theme' represents an Indo-European, perhaps 
Iranian, element in the Turk system of beliefs linking at least some sections of the Turk 
ruling class to the Sogdians and, beyond them, to the Wu-sun who - for all we know - 
may have been Iranians. At the same time, it must be emphatically stated that the widely 
accepted view according to which the Wu-sun had blue eyes and blond hair rests on a 
textual misunderstanding, as was shown by Otto Franke as early as 1904. 

Nor is the theme of the cavern a literary invention: it was a belief actively held by the 
Turks. The Yu-yang tsa-tsu even gives the name of the cavern (A-shih-te), which is that of 
a Turk clan of great importance, a perennial antagonist of the A-shih-na clan, whose claim 
to rule the Turks is clearly implied by the other two legends. The legend of 'The Spirit 
of the Lake' speaks of a 'birth cavern' and the existence of such an 'ancestral cavern' 
is demonstrated by the Chou shu's statement that every year the Turk kaghan leads the 
notables of his people 'to the ancestral cavern to offer a sacrifice'. The Tung tien (193, 



ISBN 978-92-3-103211-0 Contents Ethnogenesis 

14a) cites from a work (now, alas, lost) entitled On the Origin and Development of the 
Turks in which the caverns of the Turks were also mentioned. 

Of the various themes which can be identified in the Turk ancestral legends, those of the 
wolf and the cavern appear together only in one, namely, 'The Abandoned Child Brought 
up by a Wolf, which justifies the rule of the A-shih-na clan. The theme of the cavern also 
appears in 'The Spirit of the Lake', though this story differs so fundamentally from the 
other two legends that it cannot be ascribed to the same clan. It is just possible - and the 
suggestion is made with the utmost caution - that this theme indicates some links with the 

The composite character of the Turk nation, as revealed by an examination of their 
ancestral legends, is supported by evidence of a different nature. Turk civilization as we 
know it from the written sources contained a number of specific elements which are atyp- 
ical for Turkic peoples. Among them were, first, a system of orientation facing east; and, 
second, an unusual system of numerals where, in double-digit numbers, the tens are indi- 
cated by the next highest multiple of ten, e.g. bir otuz '21 ' (= one thirty), and which cannot 
be of Turkic origin. 

It has been established beyond doubt that the population of the Turk Empire was mul- 
tilingual - the existence of the Bugut inscription, written in Sogdian, would in itself prove 
this point. It has also been shown that Turk - as used in the inscriptions of the Orkhon 
- contains a number of Samoyed or Ugric loan words which are specific to this language 
and form no part of the common Turkic vocabulary. Even in the solemn, funerary style 
of these inscriptions, Ugric or Samoyed words appear, expressing concepts as common as 
'word' (ay, sab) or, indeed, 'horse' (yunt). Their occurrence in Turk indicates the presence 
of Ugric or Samoyed elements in that stratum of Turk society which had some cultural 
influence on the ruling class. It can be taken for granted that the language of the funerary 
inscriptions (be it Sogdian as in Bugut, or Turkic as on the Orkhon) was that of the con- 
temporary ruling class. Reference in the Orkhon inscriptions to a feminine deity, Umay, of 
clearly Mongol origin, attests to the presence of some Mongol element within the fabric of 
early Turk civilization. 

Turk personal names appear in a great variety of sources and scripts which, apart from 
Turk itself, include Chinese, Sogdian and Greek. No methodical, comprehensive study of 
these names has been undertaken, but even a cursory examination of Turk anthroponyms 
reveals a substantial number which cannot be explained from Turkic. It is seldom easy to 
reconstruct the original form of a proper name that is attested only in Chinese transcrip- 
tion. Nevertheless there have been several successful attempts, particularly in instances 
where the language to which the name belongs is known - as, for example, with Chinese 



ISBN 978-92-3-103211-0 Contents The economy 

transcriptions of Buddhist technical terms. However, the original Turkic forms of many 
Turk proper names have not yet been established. Such is the case, for example, of the clan 
names A-shih-na or A-shih-te (attempts to see in the former a Mongol (?) word meaning 
'wolf lack proper phonetic foundations). Nor has a satisfactory Turkic form been estab- 
lished for the Turk personal name that is written as Silziboulos in Greek characters. 

In several instances (for example, Nivar kaghan; see page 333) the personal names 
of Turk dignitaries are clearly non-Turkic - with the exception of a few, clearly specified 
cases, the initial n- does not occur in Old Turkic. It is well known that a Turkic word cannot 
begin with a consonant cluster, yet the names of two Turk chiefs appear in Greek sources 
as Spartseugoun and Stembis. The latter name is known also in Chinese transcription and 
occurs in the Orkhon inscriptions either with or without the initial i-\ shtmi or ishtmi, to be 
read Ishtemi. There is no reason why the Greek transcription would have ignored an initial 
i- had the name had one in its original form. There are, however, very good linguistic 
reasons for the Turks to attach a prosthetic i- to an initial s?-consonant cluster. Clearly, the 
name Ishtemi, though borne by a Turk ruler (see page 332), was not Turkic. Although a 
detailed examination of Turk proper names is beyond the scope of the present chapter, it is 
clear that many of the personal and tribal names and dignitary titles used by the Turks are 
neither Turkic, nor Mongol, nor Iranian. 

The economy 

In 552, as previously mentioned, a successful Turk uprising overthrew the Juan-juan ruler 
A-na-kui and effectively ended the Juan-juan Empire which, for the previous century and 
a half, had been the dominant power on the eastern steppe. Bumin, the leader of the coup, 
is said to have been angered by A-na-kui's refusal to grant him the hand of one of his 
daughters on the grounds that the suitor was merely a 'blacksmith slave' and thus unworthy 
of such an honour. There is overwhelming evidence that the Turks - the people of Bumin 
- were originally a group of metallurgists engaged in the mining or processing of iron, 
or possibly both. The above-mentioned 'caverns' of the Turks were, in fact, underground 
mines where they laboured for the principal benefit of the Juan-juan. 

Thus the overthrow of Juan-juan rule was not the result of an invasion by an external 
enemy but was brought about by an internal upheaval, the revolt of a discontented faction 
which, ethnically or linguistically, may not have been different from the dominant group. 
(There is as yet no conclusive evidence as to the language of the Juan-juan - it might even 
have been Turkic.) The only distinction between the Juan-juan and the Turks which can 
be established with any degree of certainty relates to occupation. To put it in simple terms, 



ISBN 978-92-3-103211-0 Contents Political history 

through Bumin's action the reins of power were seized by the metallurgists of the Juan-juan 

Although (if for no other reason than military necessity) pastoral nomadism was the 
dominant economic activity of the ruling stratum of the newly created Turk state, it did not 
involve the whole population. Besides metallurgy - which, at some time, the Turks seem 
to have left to the Kyrgyz - important sections of the population must have continued to 
provide for themselves through hunting and fishing, the traditional economic activities of 
the forest region where pastoral nomadism could not be practised and where many Turks 
continued to live. That the leaders were preoccupied with the necessities of daily life is 
evident from the words on Bilge kaghan's funeral stele: T [Bilge] did not reign over a 
people that was rich; I reigned over a people weak and frightened, a people that had no food 
in their bellies and no cloth on their backs.' In some campaigns the Turks were even short of 
horses. The inscription of Tonyuquq (see below) reveals that at least on one occasion 'two 
parts [of the Turk army] were mounted, one part was on foot' . The very precariousness 
of their existence made the Turks, or at least their leaders, vulnerable to the lure of those 
'Chinese riches' mentioned in the inscriptions. 

Political history 

Bumin died shortly after he had deposed A-na-kui. He was followed by his son Kuo-lo 
(Qara?), who ruled for only a few months. On his death, the government of the newly cre- 
ated Turk Empire was divided between Bumin's other son Muhan (553-572) and Ishtemi 
(553- ?), Muhan's uncle, and brother of the late Bumin. Muhan ruled over the eastern part 
of the empire, centred on Mongolia, while Ishtemi was in charge of the western areas. The 
heart of the empire - where the 'ancestral cavern' and Mount Otiikan, the sacred forest of 
the Turks, were located - was the eastern part. Thus it can be said that almost from the 
moment of its inception, the Turk Empire was bicephalous. 

Uncle and nephew embarked on a series of military campaigns. In the east, this brought 
victory over the Kitans and the incorporation of the Kyrgyz into the Turk state. In the west, 
between 557 and 561, the Hephthalite Empire was crushed through a joint action of the 
Turks (probably led by Ishtemi) and the Sasanian king, Khusrau I Anushirvan, resulting in 
the establishment of a common border between the two empires. Through their conquests 
the Turks now controlled large sections of the trade routes to the West and, edged by 
their Sogdian subjects, they wished to take advantage of the lucrative silk trade formerly 
dominated by the Hephthalites. After their attempts to establish commercial footholds in 
Persia met with failure, they aimed to bypass Persia altogether and establish direct links 
with Byzantium, the principal consumer of silk. 



ISBN 978-92-3-103211-0 Contents Political history 

The first Turk delegation known to us arrived in Constantinople in 563. It had been sent 
by Askel, head of the first tribe of the Nu-shih-pi tribal federation of the Western Turks. 
It was followed five years later by a more substantial trade delegation headed by a Sog- 
dian called Maniakh. He was received by Emperor Justin II, who was more interested in 
securing an ally to the rear of the Sasanians (with whom, since 527, Byzantium had been 
in almost permanent conflict) than in the importation of silk. According to the Byzantine 
historian Menander, the Turk ruler on whose behalf Maniakh negotiated was Silziboulos 
(who is usually wrongly identified with Ishtemi, even though this name is rendered Stem- 
bis in Greek sources; see above). Silziboulos and his son Turxath were minor rulers in the 
westernmost parts of the Turk Empire, perhaps on the same level of authority as the pre- 
viously mentioned Askel. Menander clearly states that Turxath was but one of the eight 
chiefs among whom rule over the Turks was divided. 

On his return journey, Maniakh was accompanied by a Byzantine counter- embassy 
led by the strategos Zemarkhos, who was, in his turn, very well received by Silziboulos. 
Other diplomatic exchanges followed until 572 when, on his second mission to the Turks, 
the Byzantine envoy Valentine was received by Turxath (perhaps Turk shad), son of the 
just deceased Silziboulos. In sign of mourning, members of the Byzantine delegation were 
not only requested to lacerate their faces, but were given a bitterly hostile reception by 
Turxath, who accused the Byzantine emperor of treason for having given asylum to the 
Avars (considered by him to be fugitive subjects of the Turks: see Chapter 13). At that 
time the principal ruler of the Western Frontier Region of the Turk Empire was Tardu, a 
son of Ishtemi, whose year of accession is unknown, although it cannot have been later 
than 572 since it was to him that the irate Turxath sent Valentine. 

The principal Turk ruler Muhan (553-572) was followed on the throne by his younger 
brother Taghpar (572-581). ! Having converted to Buddhism, Taghpar embarked on an 
ambitious programme of building monasteries and sponsoring the translation of Buddhist 
canonical works, presumably from Chinese into Sogdian and Turkic. These activities con- 
tinued under Taghpar's brother, Nivar (581-587) (his name is read Jibii by Harmatta), 
whose court became an important centre of Buddhist learning. It was at the time of Nivar 
that the rift separating Eastern and Western Turks occurred, an event which has long been 
thought to have taken place immediately following the death of Bumin. 

At Taghpar's death Muhan's son, Apa kaghan (known as Ta-lo-pien in Chinese sources), 
had not taken kindly to his uncle Nivar occupying the throne. He enlisted the help of Tardu 
but failed to oust Nivar, who received support from his father-in-law, the Sui emperor Kao- 
tsu. Obsessed with the desire to have a state of his own, Apa kaghan then turned against 

1 Harmatta's reading instead of the currently used Taspar. 



ISBN 978-92-3-103211-0 Contents Political history 

his former ally Tardu, chased him from his domain and established the state of the Western 
Turks opposed to that of the eastern parts controlled by Nivar. In 585 Tardu fled to the Sui 
court; nothing further is known of his activities until 594, when he reappeared in a conflict 
with the Eastern Turk kaghan, Yung-yii-lii (588-599). It seems likely that the Turk kaghan 
who, in 598, wrote a letter to the Byzantine emperor Maurice describing himself as 'lord 
of the seven races, master of the seven climes' was Tardu, once again riding high, ruling 
over an illdetermined part of Turk territory. The destiny of this extraordinary man remains 
unknown; chased by a revolt in 603, he fled and we lose his traces for ever. 

Apa kaghan did not live to enjoy the fruits of his treacherous victory over Tardu. After 
being taken prisoner by Nivar's successor Ch'u-lo-hu (587- 588), he disappeared from the 
stage of history. His place was taken by Ni-li (587-604?), a somewhat shadowy figure; he 
may have been the 'Great King of the Turks' who, according to the Armenian historian 
Sebeos, was killed in battle in 589 by the Sasanian general, Bahram Chobin, while fighting 
Hormizd IV in Persia. Turk involvement in Iranian affairs continued under T'ungyabghu 
(619- 630), kaghan of the Western Turks, an ally of Emperor Heraclius against Khusrau 
II. T'ung yabghu received the Chinese pilgrim Hsiian-tsang, who was duly impressed by 
the magnificence of the Turk court. However, pride and unbridled ambition caused T'ung 
yabghu's, downfall. In the words of the Tang shu, he was no longer 'good to his people 
and the tribes hated him' and he fell victim to a revolt led by the Karluks. 

The Chinese, past masters in the art of fighting Barbarians with Barbarians, exploited 
the Turks' endemic internal dissensions to the full. In the words of the Sui shu, 'The Turks 
prefer to destroy each other rather than to live sideby-side. They have 1,000, nay 10,000 
clans who are hostile and kill one another.' Because, owing to their location, the East- 
ern Turks presented the greater danger, the short-lived Chinese dynasties of the period 
attempted to keep them at bay with trade concessions which were tantamount to the pay- 
ing of tribute. Thus, for example, both Muhan and Taghpar received 100,000 pieces of silk 
per year from the Northern Chou, a gift barely compensated for by the horses sent to China 
by Taghpar. 

Both Kao-tsu, founder of the T'ang dynasty, and his son T'ai-tsung skilfully played off 
one Turk ruler against another. Constant attempts were made to persuade T'ung yabghu and 
other Western Turk rulers to keep the Eastern Turk kaghans Shih-pi (609-619) and Hsieh-li 
(619-634) at bay, but final defeat came through direct Chinese victory. T'ai-tsung's troops 
routed those of Hsieh-li who, taken prisoner, died in China. With his death, darkness would 
descend on the Eastern Turk Empire for half a century. 

The Western Turk Empire, which was less bothersome for the Chinese, was left to 
its own devices of self-destruction. Always a confederation of tribes acting more or less 



ISBN 978-92-3-103211-0 Contents Resurgence of the Turk Empire 

independently, the Western Turk state was swiftly falling apart. The Ten Arrows (On oq) 
or, as the Chinese called them, the Ten Clans were rent by the murderous conflicts of their 
leaders. The Orkhon inscriptions give the reasons for the internal decay: 

Because of discord between the nobles and the commoners, because of the cunning and 
deceitfulness of the Chinese who set against each other younger and elder brothers, nobles 
and commoners, the Turk people caused the disintegration of the empire that had been their 
own, [and] caused the ruin of the kaghan who had been their kaghan. 

Ho-lu, the last de facto ruler of the Western Turks, was captured by the Chinese in 657 
and died two years later, to be buried beside Hsieh-li. Thus the two rulers of fratricidal 
Turk Empire were put to rest, side-by-side and in Chinese soil. 

Part Two 


(S. G. Klyashtorny) 

Resurgence of the Turk Empire 

After the First Turk Empire had been defeated by the emperor T'ai-tsung in 630, the East- 
ern Turk tribes were resettled north of the Ordos and Shansi. T'ai-tsung drafted his new 
subjects into the service of the T'ang Empire, but the existing tribal and administrative sys- 
tem was not altered and measures were taken to attract the Turk aristocracy to the imperial 
service. The author of the ancient Turk inscriptions in honour of Kiil-tegin (732) notes with 
disapproval, when speaking of those times, that 'The Turk begs abandoned their Turk titles. 
The begs who went to China held Chinese titles, obeyed the Chinese emperor; they served 
him for fifty years' (KT, E 7-8). For most of the Turk people, forcibly resettled in strictly 
defined regions, life was hard. The Turk historian recalls those five decades as a time of 
shame, degradation and humiliation; the heaviest burden was the 'blood tribute', the oblig- 
ation to fight in the imperial wars: 'Your blood flowed like a river; your bones were heaped 
up like a mountain; your beg-like sons became slaves; your lady-like daughters became 
servants' (KT, E 23). 



ISBN 978-92-3-103211-0 Contents Political and social structure 

The Turk uprising in 679-681 was at first unsuccessful, although it led, in 682, to the 
withdrawal of Kutlug-chor, one of the Turk leaders of the kaghan tribe of the A-shih-na, 
into the Gobi desert. Once they had established themselves in the Yin Shan mountains 
(Cugay quz'i in ancient Turkic), Kutlugchor and his closest comrade-in-arms, Tonyuquq, 
succeeded in winning the support of most of the Turks and conducted successful mili- 
tary operations against the imperial forces in Shansi between 682 and 687. Kutlug-chor 
proclaimed himself Ilterish kaghan, and in so doing ushered in the resurgent Turk Empire. 

In 687 Ilterish kaghan left the Yin Shan mountains and turned his united and battle- 
hardened army to the conquest of the Turk heartlands in central and northern Mongolia. 
Between 687 and 691 the Tokuz-Oghuz tribes and the Uighurs, who had occupied these 
territories, were routed and subjugated; their chief, Abuz kaghan, fell in battle. The cen- 
tre of the Second Turk Empire shifted to the Otiikan mountains (now called the Khangai 
mountains), on the rivers Orkhon, Selenga and Tola. Having united two powerful tribal 
groups under his command - the Turks and the Tokuz-Oghuz - Ilterish kaghan was now a 
dangerous menace to the T'ang Empire. 2 

Political and social structure 

Under Ilterish kaghan the traditional structure of the Turk state was restored. The empire 
created by Ilterish and his successors was a territorial union of ethnically related and hier- 
archically co-ordinated tribes and tribal groups; they were ideologically linked by common 
beliefs and accepted genealogies, and politically united by a single military and adminis- 
trative organization (el) and by general legal norms (torus). The tribal organization (bodun) 
and the political structure (el) complemented one another, defining the strength and dura- 
bility of social ties; in the words of the Turk inscriptions, the khan (kaghan) 'el tutup 
boduriim basladinC (controlled the state and was head of the tribal group). 

The principal group in the empire was composed of twelve Turk tribes headed by the 
dynastic tribe of the Ashina. 3 Next in political importance was the Tokuz-Oghuz tribal 
group of 'nine Oghuz [tribes]'. 4 The Tokuz-Oghuz were more numerous than the Turk 
tribes themselves, but were politically less united; however, at the beginning of the seventh 
century, they were united under the Uighurs, themselves a group of ten tribes led by the 
dynastic tribe of the Yaghlakar. Two further confederations of tribes played an active role 
in the political life of the empire - the Karluks and the Basmils. Each individual tribe 
had its leader, the irkin, and each tribal group was headed by an elteber. Turk monuments 

2 Klyashtorny, 1964, pp. 25-32. 

3 Czegledy, 1972, pp. 275-81. 

4 Czegledy, 1982, pp. 89-93. 



ISBN 978-92-3-103211-0 Contents Political and social structure 

frequently mention these important representatives of the tribal aristocracy - the elteber of 
the Uighurs, the great irkin of the Bayirku, and others. 

The administrative structure of the empire, which incorporated the tribal leaders, was 
more complex. At the head of the administration stood the kaghan and his closest kinsmen, 
who held the titles of shad and yabghu. The kaghan was surrounded by his counsellors 
(buyur), who discharged military, administrative, diplomatic and legal functions and bore 
titles such as tarkhan, chor and tudun. In order to facilitate the administration, the tribes 
were divided into two territorial groups, the Tardush (western) and the Tolish (eastern). 
The soldiery of these two groups composed the right and left wings of the army's battle 
order, and they were led by the close kinsmen of the kaghan (the shads,) and the most 
influential tribal leaders of each wing. 

With its dual system of tribal and political principles, the administrative structure was a 
natural reflection of the social structure of the ancient Turk community. Its highest stratum 
consisted of begs (begler in Turk), a hereditary aristocracy; it was composed of members of 
families whose special status in the management of the affairs of the tribe was considered 
unchallengeable and hallowed by tradition. The dynastic families and tribes (the A-shih-na, 
the A-shih- te and the Yaghlakar) formed the elite of this hereditary aristocracy. Another 
stratum of that same community was the igil qara bodun (the 'common people'). Any 
deterioration of relations between strata or tribes represented a grave threat to the politi- 
cal organization of the empire. The kaghan, who personified the unity of the community 
and exploited its military and economic potential to the full, clearly had a vested interest 
in minimizing all opposition. As the manifestos recorded in inscriptions dating from the 
kaghanate show, there were frequent appeals for unity between the begs and the people 
and for obedience to the kaghan. 

Whereas the kaghan was the personification and supreme power of the community, its 
base was the fraternity of full male members of the family and tribe, who were designated 
er (man-warrior). Any youth could become an er when he reached a certain age and had 
accomplished an initiation rite (some exploit in battle or the hunt), receiving his er aty 
(man's or hero's name) whether he was one of the hundreds of common soldiers or a 
prince of the royal line. In practice, however, the situation of an er in the tribe and in the 
state depended on his rank and riches. 

Epigraphic and archaeological records show that there was a considerable degree of 
social and material discrimination within the Turk tribes. Wealth became a subject of pride 
and praise for the Turk aristocracy. Rich men (bays) are contrasted in Turk inscriptions 
with the poor (Cygay), who are described as 'pitiful, insignificant and base'. Far from 
arousing sympathy, poverty was despised. A real er would obtain riches by force of arms. 



ISBN 978-92-3-103211-0 Contents Relations with China 

The inscriptions often list the spoils of war - gold, silver and slaves, both male and female. 
The Turks' principal wealth and most coveted booty, however, was livestock, especially 
herds of horses. 'The Turk people were hungry. I took the cattle and fed them,' says Bilge 
kaghan in his account of one of his campaigns (BK, E 38). 

The burial mounds of the common soldiers, where the saddled warhorse lay next to its 
fully armed master, pale into insignificance when compared to the burial chambers of the 
higher aristocracy. In the graves of the poorest peasants, however, neither costly weapons 
nor horses were to be found. Impoverished nomads who had lost their livestock were settled 
in winter quarters and in small, permanent settlements (bali'qs), where they engaged in a 
primitive form of agriculture. They mainly sowed millet and built small forts (qurgans or 
kurgans) in which to store their grain. 

Some impoverished members of a tribe would maintain their nomadic way of life with 
the help of rich relations. Free ers of slender means inevitably became dependent on the 
begs, whose bodyguard and servants were drawn from their ranks. But no matter what 
quarrels soured relations between poor and rich ers, and between begs, and the 'common 
people', the community as a whole was quite distinct from another sector of the population 
- the slaves (qui kiin, male and female slaves), who were entirely dependent on the ers, 
enjoyed no rights and formed the periphery of ancient Turk society. 

The basis of the economy of the Turk tribes was nomadic cattle-raising. The organized 
hunt in the steppes and mountains was of military as well as economic significance: it 
was during such hunts that warriors were trained and the various detachments were co- 
ordinated. A Chinese chronicler describes the economy and way of life of the Turks thus: 
'They live in felt tents and wander following the water and the grass.' Horses were of vital 
importance to the Turks. Although the economy rested on cattle-raising, winter feed for 
livestock was not stored. The advantage of the horse was that it could be at grass all the 
year round, feeding even under a light cover of snow. Sheep and goats followed the horses, 
eating the grass that they themselves would have been unable to clear of snow. Bulls, yaks 
and camels are also frequently mentioned in Turk texts as valuable items of livestock. 

Relations with China 

In 691 Ilterish kaghan died and was succeeded by his younger brother, who assumed 
the title Kapagan kaghan ('Conquering kaghan'; Mo-ch'o in Chinese sources). His reign 
(691-716) marked the apogee of the military and political might of the Second Turk 
Empire - and the beginning of its decline. 



ISBN 978-92-3-103211-0 Contents The empire in crisis 

Between 693 and 706 Kapagan's army forced a crossing of the Huang He (Yellow 
River) six times and made deep inroads into northern China against which the Chinese 
forces could offer no effective resistance. The empress Wu paid vast indemnities to Kapa- 
gan and sent him gifts, which were in effect thinly disguised tributes. In 696-697 Kapagan 
subjugated the Kitan tribes and sealed an alliance with the Khi (tataby in Turkic texts), 
which stemmed the advance of the Chinese armies to the north-east, into the foot-hills of 
the Khingan, and secured the empire's eastern frontier. Between 698 and 701 the north- 
ern and western frontiers of Kapagan's state were defined by the Tannu Ola, Altai and 
Tarbagatai mountain ranges. After defeating the Bayirku tribe in 706-707, the Turks occu- 
pied lands extending from the upper reaches of the Kerulen to Lake Baikal. In 709-710 
the Turk forces subjugated the Az and the Chik (tribes living in Tannu Tuva), crossed the 
Sayan mountains (the Kogmen yis in Turkic texts), and inflicted a crushing defeat on the 
Yenisey Kyrgyz. The Kyrgyz ruler, Bars beg, fell in battle; his descendants were to remain 
vassals of the 'kaghan of the Otiikan mountains' for several generations. 

In 71 1 the Turk forces, led by Tonyuquq, crossed the Mongolian Altai, clashed with the 
Tiirgesh army in Dzungaria, on the River Boluchu, and won an outright victory. Tonyuquq 
forced a crossing over the Syr Darya in pursuit of the retreating Tiirgesh, leading his troops 
to the border of Tokharistan. However, in battles with the Arabs near Samarkand the Turk 
forces were cut off from their rear services and suffered considerable losses; they had 
difficulty in returning to the Altai in 713-714. There they reinforced the army that was 
preparing to besiege Beshbalyk (Pei-t'ing). The siege was unsuccessful and, after losing in 
six skirmishes, the Turks raised it. 5 

The empire in crisis 

These military defeats changed the situation drastically, serving as a signal for formerly 
submissive tribes to revolt. The Kitans and the Khi seceded; and first the Karluks then all 
the Tokuz-Oghuz tribes revolted, the latter representing a particularly serious threat to the 
Turk Empire. The Toguz-Oghuz were defeated in five battles in 715, but the revolt was 
not crushed. The following year, the great irkin of the Bayirku tribes fell on Kapagan's 
headquarters on the Tola river. Although the attack was repulsed, Kapagan himself was 
ambushed and killed. 

Kapagan had tried to change the existing order of succession, according to which he 
would be succeeded by the elder son of Ilterish, known as 'the shad of the Tardush' since 
698, who in turn would be followed by his younger brother Kiil-tegin. At the time of 

5 Klyashtorny, 1964, pp. 35-40. 



ISBN 978-92-3-10321 1-0 Contents The last war with Tang China 

Kapagan's death, both were renowned generals. Nevertheless, with the help of Kapagan's 
retainers, his son Bogii came to power. After this flagrant breaking of the law, Kiil-tegin, 
who was the hero of many battles and very popular with the forces, and enjoyed the support 
of all the influential Turk families, attacked the headquarters. He killed Bogii kaghan and 
many of Kapagan's retainers, and then set on the throne his elder brother, known as Bilge 
kaghan ('Wise kaghan'), who ruled from 716 to 734. 

Bilge kaghan mounted the throne at a time when the empire founded by his father 
was on the verge of collapse. The western lands seceded for good; and immediately after 
the death of Kapagan, the Tiirgesh leader Suluk proclaimed himself kaghan. The Kitan 
and Tatabi tribes refused to pay tribute; the Oghuz revolt continued; and the Tiirk tribes 
themselves began to rebel. Feeling unable to control the situation, Bilge kaghan offered the 
throne to his brother, Kiil-tegin. The latter, however, would not go against the legal order 
of succession. Then, at last, Bilge decided to act. Kiil-tegin was put at the head of the army, 
and the septuagenarian Tonyuquq, who enjoyed great authority among the tribes, became 
the kaghan' 's closest adviser. Bilge and Kiil-tegin now attacked the Uighurs; the rout of the 
Uighurs broke the resistance of the Tokuz-Oghuz tribes and the rich spoils heartened the 
Tiirk forces. In the summer of 718 Bilge crushed the Tatabi and the Kitans and regained 
possession of the Khingan. The detachment led by Tudun Yamtar, one of Bilge's captains, 
attacked the Karluk tribes, forced them to submit and took vast herds of horses, which were 
distributed among the tribes loyal to Bilge. 

In 7 1 8 those Tiirk and Oghuz tribes which had fled to China during the time of internecine 
strife in 716 returned to Bilge's empire. The kaghan's army became so strong that he 
decided to resume the war with China, his southern neighbour, which had offered help and 
protection to his adversaries. In the face of determined opposition from his chief counsellor, 
Tonyuquq, however, Bilge sent an embassy to Ch'ang-an instead, proposing a peace treaty. 
The emperor Hsiian-tsung, who had pacified his border with Tibet, refused to negotiate 
with the Tiirks in the hope of destroying their state, which had nevertheless been weakened 
by the internecine strife. 

The last war with T'ang China 

In 720 the Chinese army, whose main attacking force was the cavalry of its confederates - 
the Basmil, Kitan and Tatabi tribes - advanced on the Otiikan mountains in two directions. 
Tonyuquq's army met the Basmils and defeated them, taking Beshbalyk as they pursued the 
defeated tribe. The Tiirks then swiftly invaded Gansu and annihilated the Chinese garrison 
in the vicinity of Liang chou. The army of Bilge kaghan, which had participated in the 



ISBN 978-92-3-103211-0 Contents The final decade 

victory over the Basmils and in the Gansu campaign, wheeled east and inflicted further 
severe defeats on the Kitans and Tatabi. In 721 Hsiian-tsung immediately accepted the new 
peace proposals. 

The war of 720-721 was the last between the Second Turk Empire and China. The 
Chinese emperor no longer dared to disturb the peace, and Bilge kaghan broke with the 
policy of his predecessor, Kapagan. He strove consistently to expand trade with China and 
to establish family ties with the imperial House of Li, conserving the territorial status quo 
and the policy of non-interference in each other's internal affairs. It should be noted that 
Hsiian-tsung paid dearly for peace on his northern frontier, a fact that Bilge kaghan does not 
fail to mention in Kiil-tegin's epitaph: T made peace with the Chinese people; they gave us 
gold, silver and silk in abundance' (KT, S 4-5). In the year 727 alone, the Chinese emperor 
gave Bilge kaghan a 'present' of 100,000 pieces of silk in return for a symbolic 'tribute' 
of 30 horses. And it was not until 734 that the Chinese participated in a war between the 
Kitans and the Tatabi, siding with the latter; Bilge kaghan, fearing for his eastern frontier, 
fought against the Tatabi and defeated them. There were no direct confrontations between 
Chinese and Turk forces. 

The winter of 723-724 was a hard one for the Turks: they lost most of their cattle 
because of the icy conditions. In spring the war with the Oghuz and the Tatars broke out 
afresh, imperilling the gains of previous years. It took a supreme effort to defeat the rebels 
and ensure political stability within the kaghanate. In 727 Bilge kaghan refused to ally him- 
self with the Tibetans against China; he was rewarded with concessions from the imperial 
government in Ch'ang-an, allowing an expansion of the frontier trade between the nomads 
and Chinese merchants. From then on, gifts of huge amounts of silk arrived annually. 

In 732 Bilge kaghan entered the sixteenth year of his reign. 'By the grace of Heaven 
and because of good fortune and propitious circumstances, I brought back to life the dying 
people, the naked people I clothed, and I made the few many' (KT, E 29). 6 

The final decade 

Kiil-tegin died in 73 1 . Bilge kaghan did not long outlive his brother - in 734 he was poi- 
soned by one of his retainers. Near the River Orkhon, in the Koshotsaidam basin between 
the mountains, monuments were erected to both brothers with inscriptions that chronicled 
the turbulent history of the Second Turk Empire (see below). 

Bilge's successors, his short-lived sons Izhan kaghan (734-739) and Tengri kaghan 
(740-741), did not depart from their father's policies, but with Tengri kaghan 's death, in 

6 English translation: Sinor, 1990, p. 313. 



ISBN 978-92-3-103211-0 


Epigraphic memorials of the Turks 

741, the empire began to disintegrate. Apanaged rulers of the kaghan's House of Ashina 
were less and less able to cope with central power. The young Tengri kaghan was killed 
by his uncle, Kutlug yabghu, who seized power. War broke out with the tribal groups of 
the Uighurs, the Basmils and the Karluks, and Kutlug yabghu and his followers perished in 
the fighting. In 745 the Second Turk Empire ceased to exist. The Turk tribes, who retained 
part of their lands, played no significant role in succeeding events. The last reference to 
them in Chinese sources relates to the year 94 1. 7 

Epigraphic memorials of the Turks 

Turk stone inscriptions date back to the second half of the sixth century, when a stele 
in honour of the Turk kaghan Taspar (Taghpar in Harmatta's reading) was erected, with 
inscriptions in Sogdian and Sanskrit (the Bugut inscription, 582) (Figs. 1 and 2). 8 However, 
the heyday of ancient Turk epigraphy - in the original runic script, which was invented no 

FIG. 1. Mongolia. Stele with the Bugut inscription. (Photo: © S. G. Klyashtorny.) 

7 Klyashtorny, 1964, pp. 41-3. 

8 Klyashtorny and Livshits, 1972, pp. 69-102. 



ISBN 978-92-3-10321 1-0 Contents Epigraphic memorials of the Turks 

, I 

FIG. 2. Mongolia. Stele with the Bugut inscription. (Photo: © S. G. Klyashtorny.) 

FIG. 3. Mongolia. Bulgan aimak. Obverse of the stele. (Photo: © S. G. Klyashtorny.) 



ISBN 978-92-3-103211-0 


Epigraphic memorials of the Turks 

FIG. 4. Mongolia. Bulgan aimak. Reverse of the stele. (Photo: © S. G. Klyashtorny.) 

-* Mm 

w ^ 

" ' * ' 4 

ftSK 7 ^' # ' 

FIG. 5. Mongolia. Gobi-Altai aimak. (Photo: © S. G. Klyashtorny.) 



ISBN 978-92-3-103211-0 


Epigraphic memorials of the Turks 


FIG. 6. Mongolia. Gobi-Altai aimak. (Photo: © S. G. Klyashtorny.) 

later than the middle of the seventh century, and in the Turk language - was the era of 
the Second Turk Empire. The earliest example appears to be the memorial from Choiren, 
which dates from between 687 and 691; the inscription tells of the Turks' return to their 
lands in northern Mongolia and of their victory over the Tokuz-Oghuz. The largest and 
most significant memorials are the Koshotsaidam steles (monuments in honour of Kiil- 
tegin and Bilge kaghan, who died in 73 1 and 734 respectively), written on behalf of Bilge 
kaghan by his nephew Yollyg-tegin; and the memorial written by Tonyuquq after 716 and 
subsequently incorporated in his burial mound. 

Further inscriptions of this kind are known to us; these historical and biographical texts 
are memorials or eulogies for the living, and they tell of the deeds of Turk kaghans and their 
retainers. They combine descriptions of events that involved the hero of the inscription (or 
his ancestors) with an exposition of the political beliefs and ideas of the author of the text; 



ISBN 978-92-3-103211-0 Contents The Turgesh state 

they may be seen as 'declarations of intent' and to some extent were used as propaganda 
(Figs. 3 and 4). Even more common were memorial inscriptions on rock faces, some of 
which proclaimed the author's right to use the adjacent pasture or site (Figs. 5 and 6). 9 

The Turgesh state 

At the end of the seventh century Wu-chih-le (699-706), leader of the Turgesh tribes that 
lived in the western T'ien Shan mountains, had driven the T'ang protege Bori-shad out 
of Semirechye and established his own power over the territory from Chach (present-day 
Tashkent) to Dzungaria. 'Major' and 'minor' headquarters for the kaghan were established 
in Nevaket on the rivers Chu and Hi, and the country was divided into 20 districts ruled 
by the kaghan's stewards (tutuk), each of whom was able to muster between 5,000 and 
7,000 warriors. Wu-chih-le assumed the traditional title of the Western Turk states, 'kaghan 
of the People of Ten Arrows'; and also the new title, ' Tiirgeshfczg/zan' - copper coins 
were now minted in Nevaket bearing this legend in Sogdian. Wu-chih-le's successor, Sakal 
kaghan (706-711), met with opposition from tribal leaders who supported his younger 
brother's claim to the throne. The Eastern Turk leader, Kapagan kaghan, intervened in the 
civil unrest, and after the defeat of the Turgesh forces in a battle on the River Boluchu 
in 711 both brothers were killed. What was left of the Turgesh army was rallied by the 
commander, Suluk Chabish-chor (Su-lu in Chinese sources), and retreated beyond the Syr 
Darya into Tokharistan. It was not until 715 that Suluk, having proclaimed himself Turgesh 
kaghan, was able to return and restore the independence of his state. 

Throughout his reign (715-738) Suluk Chabish-chor had to fight on two fronts. From 
East Turkestan he was threatened by pretenders to the throne who belonged to the family 
of the Western Turk kagham and were supported by Chinese troops. Suluk obviated this 
danger by means of diplomacy (marriage to the daughter of one of the pretenders) and 
military action (laying siege to the capital of the T'ang governor of Kucha in 726-727). By 
marrying the daughters of the Eastern Turk leader Bilge kaghan and the king of Tibet, the 
Turgesh kaghan firmly consolidated his eastern flank. 

From the west the Turgesh were threatened by the conquering Arab armies, who crossed 
the Syr Darya (Jaxartes) several times in 714-715. This compelled Suluk to join battle 
with the Arabs, along with other Central Asian states striving to retain their independence. 
In 720-721 Suluk's general, Kuli-chor (Kursul in Arab sources), led successful military 
actions against the Arabs in Sogdiana. In 728-729 Suluk supported the anti-Arab revolts 
of the citizenry of Samarkand and Bukhara, and drove the Arabs from Sogdiana. It was not 

9 Klyashtorny, 1975, pp. 119-28. 



ISBN 978-92-3-10321 1-0 Contents The Uighurs and the Karluks 

until 732 that the Arab governor defeated the Tiirgesh, near Tavavis, and entered Bukhara. 
In 737 the Tiirgesh crushed the Arabs in Tokharistan, but were subsequently defeated. 
When Suluk returned to Nevaket the following year, he was killed by one of his retainers; 
and in 739 the Arabs captured and executed Kuli-chor. 

The death of Suluk and the brief reign of his son, Kut-chor kaghan (T'uho- hsien in 
Chinese sources) (738-739), marked the beginning of a 20-year struggle for power between 
the leading members of the 'yellow' and 'black' tribes, which polarized the Tiirgesh tribal 
group. Taking advantage of the internecine strife in 748, the T'ang governor of Kucha led 
an invasion force into Semirechye, capturing Suyab, one of its most important towns. In 
751, however, the Chinese forces were defeated by the Arabs and the Turks near Talas, 
and they fled Semirechye. The kaghans of the 'black' Tiirgesh seized power (749-753) but 
they were unable to end the internal conflicts. In 766 the Karluks, who had consolidated 
their hold on Semirechye after being driven from Mongolia in 746-747 by the Uighurs, 
killed the warring Tiirgesh kaghans, and the Karluk yabghu became the founder of a new 
state in the T'ien Shan mountains. 10 

The Uighurs and the Karluks 

The fall of the Second Turk Empire and the revolt of the Uighur, Basmil and Karluk tribes 
had created a political vacuum in the steppe. The struggle to obtain power and the title 
of kaghan, and to set up a new state, intensified between the Basmils (whose leader was 
proclaimed khan) and the Uighurs. The Uighurs emerged victorious. Their allies, the Kar- 
luks, who had gained no advantage from their participation in the war with the Turks, then 
allied themselves with the Tiirgesh, but, failing to defeat the Uighurs in battle, fled to the 
T'ien Shan mountains in 746. The following year, the Karluks, with the support of the Tatar 
tribes living in eastern Mongolia, attacked the Uighurs once again, but without success. In 
744, with the Uighur tribes' proclamation of Kiil-bilge kaghan as their leader, the Yagh- 
lakar dynasty came to power in the steppe. Meanwhile in Central Asia a powerful state of 
Turkic-speaking nomads - that of the Uighurs - came into being. 11 

Chavannes, 1903, pp. 279-303. 
Klyashtorny, 1982, pp. 335-66. 



ISBN 978-92-3-103211-0 Contents Three great empires 




Mu Shun-ying and Wang Yao 


Three great empires 343 

The Western Regions under the Early T'ang 344 

The kingdom of Tibet 351 

Three great empires 

The seventh century is of particular significance for the political history of Central Asia. 
In its first quarter it witnessed the rise of three great powers, each of which was to have a 
major impact on the course of events in the two succeeding centuries. In the east, China 
came under one of its most powerful and prosperous dynasties, the T'ang (618-907). At the 
foot of the towering peaks of the Himalayas, the ancient Tibetan people, the Bod (known 
as the Tu-po or more commonly as the Tu-fan in Chinese records, and as Tiipiit in Sog- 
dian and Turkic), emerged victorious from their age-old inter-tribal rivalry and were quick 
to establish a unified monarchy of the Yar-klung-spurgyal family. In the west a series of 
historical events led to the sudden rise of the Arabs. 

By a curious coincidence, all three empires were founded almost simultaneously. In the 
640s the T'ang came to dominate the oasis states of the Tarim. From the 660s onwards, 
Tibet, having established itself firmly in the Koko Nor area (now Qinghai province in 

See Map 6. 



ISBN 978-92-3-10321 1-0 Contents The Western Regions under the Early Tang 

China), began to dispute supremacy over the Gansu corridor and the Tarim with the T'ang. 
This rivalry, which lasted more than two centuries, brought several other peoples in this 
area into the conflict. Thus, a multilateral relationship between the T'ang, the Tibetans 
and many Turk confederations evolved in the Western Regions (Hsi-yii), which reflected a 
motley combination of interests. At the same time, the Arabs accomplished their conquest 
of the Sasanian Empire in 651 and continued to push eastwards. 

One hundred years later, in 751, the Arabs under the command of Ziyad b. Salih, in 
alliance with the Karluks and other Turkic peoples, defeated the Chinese forces under the 
military governor Kao Hsien-chih near the Talas river. In itself this battle was not of great 
importance, but the outbreak of the An Lushan rebellion (755-763) in China proper, and 
the consequent withdrawal of the Chinese armies from the Western Regions, left the Arabs 
in a strong position to extend their influence in Central Asia. However, according to the 
evidence of both Chinese and Tibetan sources, any further Arab advance seems to have 
been temporarily checked by the Tibetans. 1 

The Western Regions under the Early T'ang 

Before the establishment of the T'ang dynasty in 618, the Eastern and Western Turks dom- 
inated the vast expanse of the Central Asian hinterland. The Eastern Turk kaghan Hsieh-li 
(619-634) made repeated incursions into China and even reached the vicinity of Ch'ang- 
an, the T'ang capital, in 626; but the tide of events then began to turn in favour of the 
T'ang. In 630 Hsieh-li was taken prisoner by the T'ang expeditionary forces, thus ending 
the power of the Eastern Turk Kaghanate for over half a century. In the same year, the 
Western Turk kaghan, T'ung yabghu (619-630), was murdered. His death triggered off a 
series of quarrels and bitter rivalries among the Western Turks. Most of the nomadic tribes 
on the steppe, together with the oasis city-states of the Tarim basin, were caught in the con- 
flict between the T'ang and the Western Turks, changing their allegiance as the situation 

It is no accident that the first oasis city-state that the emperor T'ai-tsung (627-649), 
the real founder of the T'ang dynasty, intended to conquer was Kocho: located in the 
Turfan depression, a strategic location on the Silk Route, this oasis was the closest to 
Ch'ang-an as well as the largest of the Western Regions. Ever since the first century B.C., 
it had been more influenced by Chinese culture than other oasis states. From 502 to 640, it 
was even ruled by the royal House of Ch'ii, of Chinese origin. In 638 the king of Kocho, 
an ally of the Western Turks, was encouraged by them to defy the T'ang. The emperor 

1 Beckwith, 1980, pp. 30-8; 1987; Bacot et al., 1940^46; Chang, 1959-60. 



ISBN 978-92-3-10321 1-0 Contents The Western Regions under the Early Tang 

T'ai-tsung dispatched an expeditionary army against the king. The Western Turks, whose 
troops were stationed at Beshbalyk (present-day Jimsa in Xinjiang), an important city to 
the north of Turfan, had promised to assist Kocho in case of attack, but fled upon the 
arrival of the Chinese. In 640 Kocho was forced to surrender. The occupation of Kocho 
not only inaugurated the Chinese penetration of the Western Regions, but also increased 
tension between the Chinese and the Western Turks. The T'ang set up Hsi-chou Prefecture 
in Kocho and T'ing-chou Prefecture in Beshbalyk in 640. 

T'ang policy after the conquest of Kocho was aimed at dominating all the Western 
Regions. In spite of opposition by the Western Turks, the king of Karashahr acknowledged 
T'ang rule in 644. Two years later, Haripuspa, who had succeeded Suvarnadeva (624-646) 
as ruler of Kucha, submitted to the Western Turk Kaghanate. In response, the T'ang army 
crossed the T'ien Shan and captured the city. Most of the other oasis states hastened to offer 
their submission. On the death of Emperor T'ai-tsung in 649, Ho-lu, kaghan of the Western 
Turks, sought to reassert his supremacy over the Western Regions. After a campaign lasting 
seven years (651-657), the T'ang defeated Ho-lu and the Western Turks ceased to exist as 
a political force. 


To rule the newly conquered areas, the T'ang created a double administrative system, which 
took account of local conditions. Three oases in the eastern part of the region, where Chi- 
nese influence was most evident - Hami, Turfan and Jimsa - were incorporated into the 
Chinese civil administration. The discovery of many administrative documents bearing 
official seals has provided concrete evidence of T'ang rule in the Turfan district. 2 Nomadic 
tribes and city-states that had pledged allegiance to the T'ang court were allowed to main- 
tain their privileged status in their localities in accordance with a system called chi-mi. This 
was an institution by which the conquered rulers were 'restrained' (chi), and 'won over' 
(mi) by the honour of being invested with a title and the conferment of an embroidered silk 
robe, together with the standard, drums and horns as the emblems of mandate. 

The chi-mi system was an important means of resolving conflicts between the T'ang 
court and the various nationalities under their rule. The T'ang usually appointed repre- 
sentative members of those ruling clans, or hereditary royal families, who had pledged 
allegiance to government posts and bestowed on them honourable ranks or titles. In prin- 
ciple, all the original rulers within the chi-mi governorates or prefectures were allowed 
to continue to reside in and oversee the domestic affairs of their districts. For example, 

2 2. Tu-lu-fan chu-tu wen-shu, 1985-92. 



ISBN 978-92-3-10321 1-0 Contents The Western Regions under the Early Tang 

the T'ang set up for the nomadic Western Turks two chi-mi Protectorates-General - those 
of Meng-chih and K'un-ling - with the subordinate Turk chieftains appointed as khans, 
and protectors-general. Four chi-mi governorates were also established for the sedentary 
inhabitants of the oasis states. 

The officials in charge of the various chi-mi administrations were all members of the 
local minority elites. For example, King Su-chi (son of Haripuspa) and King Fu-shih- 
hsiung were nominated governors-general of Kucha and of Pi-sha (Visha in Khotanese, 
Vijaya in Tibetan) in Khotan respectively. Chi-mi administrators had to pledge allegiance 
to the T'ang, accepting the patents issued to them by their Chinese overlords. Their task 
was to guarantee the security of the empire's borders, while the T'ang government sta- 
tioned troops only in key locations, namely the ' Four Garrisons'. All chi-mi administrators 
regularly had to pay taxes to the T'ang court. The tribute included horses, sheep, camels, 
eagles, the skins of leopards and martens, rare birds, jade, agate, pearls, shui-ching ('germ 
of water' crystal), gold and silver wares and various kinds of woollen blankets. Of all the 
items of tribute, the horses from the Western Regions were the most important. 

To control the chi-mi governorates the T'ang set up a combined civil and military 
administration, the Protectorate-General of An-hsi (i.e. 'Pacifying the West'). It was first 
established immediately after the conquest of Kocho in 640 and moved to Kucha after the 
suppression of the revolt of Ho-lu in 657. The Four Garrisons - Kucha, Su-le (Kashgar), 
Khotan and Yen-chih (Agni in Tokharian, Karashahr in Turkic) - came under the juris- 
diction of this Protectorate- General. From 679 to 719 Suyab was listed among the Four 
Garrisons instead of Yen-chih. It should be noted that the An-hsi Protectorate-General, with 
its Four Garrisons, was in control of all the region's military and administrative affairs. In 
702 the Protectorate-General of Pei-t'ing was established at Beshbalyk to strengthen the 
T'ang position to the north of the T'ien Shan (the Celestial Mountains). 


After 659 the T'ang faced their most powerful opponent in the Tarim when the Tibetans 
appeared on the stage to challenge T'ang supremacy. Allying themselves with the revived 
Eastern Turks and various tribes of the resurgent Western Turks, the Tibetans invaded the 
Tarim and repeatedly occupied the oasis states. Under their attacks, after 663, the Chinese 
were more than once forced to withdraw and to abandon An-hsi with its Four Garrisons 
to the Tibetans. At the peak of its power, the territory controlled by Tibet ranged from the 
T'ien Shan in the north to T'ien-chu (the Chinese name for present-day India) in the south, 
and from the present western Gansu province and Sichuan in the east to eastern Central 
Asia. In 692 the T'ang finally defeated the Tibetans, restored the Protectorate-General of 



ISBN 978-92-3-10321 1-0 Contents The Western Regions under the Early Tang 

An-hsi in Kucha and recovered the Four Garrisons. At this time a Chinese force of some 
30,000 men was stationed in the Western Regions. The presence of a permanent garrison 
indicates the threat posed by the Tibetans and the T'ang court's need to keep control of the 
Four Garrisons and the route by which they were supplied. 


Under the Early T'ang, the Turkic tribes had a nomadic economy with the corresponding 
customs and way of life. Later they became semi-nomadic, transmigrating between winter 
quarters and seasonal pastures. 

Under T'ang rule, the sedentary population in the area south of the T'ien Shan adopted 
new customs. In Kucha, for example, people had their hair cut short around the top of 
their heads - the only exception was the king, who, moreover, wore a hat and robe made 
of brocade, and a bejewelled belt. At the start of each year, there were games involving 
goats, horses and fighting camels. These festivities lasted for seven days, with predictions 
about the year's harvests based on the results of such fights. The 'sprinkling with cold 
water' was a winter solstice dance. In the tenth month of the year, participants had to wear 
masks, paint their faces to resemble animals or disguise themselves as ghosts, leaping 
about to the clamour of drums and other musical instruments. They splashed cold water 
on each other and over passers-by in order to drive out devils. The local population were 
adherents of Buddhism (see pages 364-5), especially the Hinayana school. Every year, on 
the occasion of certain Buddhist festivals, people would gather to hear the exposition of 
Buddhist doctrines; statues of the Buddha were taken from the monasteries and carried on 
a 'parade of Buddhas', often drawing thousands of participants. 

Of all the cultural activities in Kucha, music and dance seem to have had the greatest 
importance and influence and to have enjoyed the greatest popularity among the people. 
The music and dance of Kucha were celebrated by many T'ang poets. 'Rainbow Skirt and 
Feathered Dress', a song and dance of Central Asian origin, was probably brought into 
Ch'ang-an, the T'ang capital, by way of Kucha. The best known of all was the popular 
music of Western Liang, a town of the Gansu corridor (now Wu-wei in Gansu province): it 
was actually an amalgam of the music of Kucha with traditional Chinese music. After being 
introduced into the two T'ang capitals, Ch'ang-an and Lo-yang, the songs and dances of the 
population of the Western Regions also became very popular. Even the so-called Ta-yileh 
(Grand Music), played at the imperial court, was mixed with Kuchean music and used 
Kuchean musical instruments, as recorded in the T'ang shu. The Kuchean four-stringed 
bent-neck lute, the oboe, the flute and the drum were among the most popular instruments 
adopted by Kuchean musicians. 



ISBN 978-92-3-10321 1-0 Contents The Western Regions under the Early Tang 

Under the influence of the musical styles of Kucha and that of Sogdiana, Chinese T'ang 
music began to sound like that of the city-states of Central Asia. Following the fashion, 
many members of the imperial family and of the aristocracy took to playing drums of the 
type widely used in musical performances in the Western Regions. During that time, almost 
all the famous musicians and dancers in the T'ang capital - for example, the Kuchean musi- 
cian Po Ming-ta - were of Central Asian origin. As for the dancing girls, with their long 
hair, fluttering sleeves and gauzy scarves, they excelled in beautiful dances with whirling 
gyrations 'as swift as the wind'. 

In the kingdom of Khotan, the people were pious and enthusiastic Buddhists and it 
seems likely that Khotan had already adopted Buddhism some time before the first century. 
Hundreds of sarighdrdmas (monasteries), studying the Mahayana doctrine, were active 
centres of religious and literary life, engaged in copying Buddhist Sanskrit and Khotanese 
Saka manuscripts, translating and adapting Indian poetry and religious literature, and so 
on. Khotan was always attractive to Chinese pilgrims seeking Buddhist scripts. 

The Khotanese paid particular attention to etiquette. Whenever they sent a letter, they 
would hold it over their heads to show respect for the recipient. They used wood to make 
pens, and jade to make seals. The people of Khotan excelled in woollen rugs and carpets 
and silk tapestries of fine workmanship, with figurative and floral motifs. They were also, 
like the people of Kucha, devoted to singing and dancing. 

Many areas of Central Asia maintained their own distinctive styles and considerably 
influenced the development of T'ang culture. The Barbarian (Hu) styles constituted a 
prominent element in T'ang culture and art; and Barbarian and Turkic costumes became 
fashionable in Ch'ang-an and Lo-yang, with even the imperial family and the aristocracy 
adopting the fashion. Many well-known painters came from the Western Regions. Among 
the painters of the seventh century, a father and son from the Yii-chi family - Yii-chi Pa- 
chih-na and Yiich'ih I-sen (Visha Irasanga in Khotanese) - were the best known. Famous 
as 'Yii-chi the Elder' and 'Yii-chi the Younger, they were members of the Khotan royal 
family. Yii-chi the Younger specialized in painting Buddhas and foreigners; a painting of a 
devaraja by this master has survived. 3 


The frequent passage of caravans along the Silk Route had not only stimulated trade, 
but also promoted the exchange between East and West of scientific, artistic and cultural 
achievements, as documented by abundant archaeological evidence. The most important 

3 Liu, 1969; Samolin, 1964; Tikhvinskiy and Litvinsky, 1988; Litvinsky (ed.), 1992. 



ISBN 978-92-3-10321 1-0 Contents The Western Regions under the Early Tang 

material remains found in the tombs at Astana and Karakhoja (two cemeteries in the sub- 
urbs of Kocho) are Chinese documents, silk fabrics, clay sculptures, wooden figures, paint- 
ings, pottery and woodwork, building equipment, dried fruit and dried foods, Byzantine 
gold and Persian silver coins, as well as T'ang copper coins. Although most of these relics 
are damaged and incomplete, they are authentic products of the T'ang period. In the paint- 
ings on silk, for example, the 'plump women' are typical of the peak period of T'ang art. 
Many of the Confucian classics were found in the tombs around Kocho. 

Archaeological evidence shows that elaborate funerals were held for the dead. The 
tombs of the period usually had one or more chambers, doors and paved approaches. In 
the case of government officials and persons of importance, the funerary objects included 
wooden models of chambers and pavilions, clay figurines representing male and female 
attendants, and men and women riding horses or playing polo. The same group of graves 
(i.e. tombs at Astana and Karakhoja) also yielded painted clay tomb-guardians or guardian 
genii, and figures showing the fabulous creatures of composite monsters with both human 
and animal features. A T'ang copper coin, a Persian silver coin or a Byzantine gold coin 
was frequently placed in the mouth of the deceased and the face was covered with a special 
type of brocade called fu-mien. 

From both the archaeological evidence and the literary sources, it is clear that there were 
workshops in the oasis states specializing in the production and manufacture of textiles, 
paper, wine, pottery, wooden objects, utensils, coaches and horse trappings. There were 
also grain-depots, firewood merchants and jewellers' shops. In the markets, people traded 
in silks, cattle, horses, camels and slaves. A new wine-making grape, the famous 'mare 
teat' or 'horsenipple' grape, was introduced from Kocho to China, and with it, knowledge 
of the art of making grape wine in 'eight colours' of this highly aromatic beverage. Another 
novelty was cotton cloth (tie-pu), spun and woven by the natives of Turfan. 

Silk had long been China's traditional export. Chinese documents unearthed from ancient 
tombs in Turfan show that Chinese silk was then available in all Central Asian markets. 
Under the T'ang, the silk industry made important progress and textile centres began to 
appear in the Western Regions. There were special brocades from Kucha, Kashgar and 
Kocho and Persian dibadj-brocade was also on sale. 4 A batch of T'ang silk products, dis- 
covered in tombs at Turfan, includes brocade, damask, lute-strings, silk gauze, and printed 
and embroidered silk fabrics. The bright colours and unique designs of this superb collec- 
tion are very impressive. Advanced Chinese techniques in textileproduction, paper-making 
and printing were transferred to Western Asia and eventually reached Europe. 

4 Tu-lu-fan chu-tu wen-shu, 1985-90, Vol. 1, pp. 181, 187; Vol. 2, pp. 18, 60. 



ISBN 978-92-3-10321 1-0 Contents The Western Regions under the Early Tang 

The art of paper manufacture was first introduced from China to Kocho (Turfan), and 
Chinese paper was imported to Samarkand as early as 650. Among the T'ang soldiers cap- 
tured by the Arabs at the battle of Talas in 75 1 were craftsmen skilled in the manufacture 
of paper, textiles, and gold and silver ornaments. 5 According to the Arabic sources, the 
paper-makers among the Chinese prisoners were taken to Samarkand to start a local pro- 
duction. Samarkand became a centre for paper-making after the eighth century and thence 
the paper industry passed to Baghdad. 6 


The dominant religion of the period was Buddhism. From the third to the ninth century, a 
vigorous Buddhist civilization developed in all the ancient oasis kingdoms, such as Khotan, 

FIG. 1. Kyzyl. Wall painting. Photo: © R.M.N./© Daniel Arnaudet. 

5 Du Huan, 'Travels', in Tu Yu, 1988, chapter on the Frontier Defence. On this work, see Rotours, des, 
1932, pp. 84, 99, 149. 

6 Maillard, 1973; 1983; Litvinsky (ed.), 1992; Yaldiz, 1987. 



ISBN 978-92-3-103211-0 Contents The kingdom of Tibet 

Kucha, Agni and Kocho. Buddhist monasteries and temples were built with donations from 
members of the local royal family, aristocrats, government officials and other rich people. 
The kings of the various oasis states became ardent patrons of Buddhism (see Chapter 18). 

During the 630s and 640s, the Chinese Buddhist pilgrim Hsiian-tsang made the round 
trip to India by way of the Western Regions. In the account of his travels, he notes that 
in the areas south of the T'ien Shan, some places have 'a dozen or so chia-lans and about 
2,000 monks', while other places have 'more than 100 chia-lam and over 5,000 monks'. 7 
(Chia-lan is a Chinese transliteration of the Sanskrit word sanghdrama, meaning 'Buddhist 
monastery'.) Hsiian-tsang also remarks that Hinayana Buddhism is prevalent in Kucha, 
while the Mahayana is represented principally in Yarkand and Khotan. 

Buddhism was also widespread among the Turks. The Western Turk kaghan, T'ung 
yabghu, was converted by an Indian monk called Prabhakaramitra. Several other minor 
Turk rulers also showed respect for, and devotion to, Buddhism. 

Most Buddhist monasteries and temples in the Western Regions were built along streams 
in mountain valleys, or in cave-temples. The best-known of these are the Kyzyl in present- 
day Bay, the Kumtura, the Kyzyl-kargha and Simsim cave-temples in the Kucha area and 
the Bezeklyk and Toyoq cave-temples in Turfan. In addition, there are large Buddhist tem- 
ples and stupas built on flat ground, such as the ruins of the Su-bashi monastery in Kucha, 
the large monastery of Miran, and Ming-oi at Karashahr, and monasteries at Kocho and 
Yar in the Turfan area. More than 1,000 Buddhist caves have been located in presentday 
Xinjiang region, most of them with painted murals. For example, the walls and ceilings 
of every cave in the Kyzyl cave-temple have murals illustrating Buddhist legends and rep- 
resenting the finest expression of Graeco-Irano- Gandharan art (Fig. 1). There are also 
styles typical of T'ang painting in the murals of the Kumtura and Bezeklyk cave-temples. 
Some temples had clay figures of Buddhas and Bodhisattvas (Fig. 2), but most are badly 

The kingdom of Tibet 

As mentioned above, a Tibetan royal line sprang up in the Yarklung valley in the first 
quarter of the seventh century. The power of this family - the Yarklung- spu-rgyal - 
spread rapidly to the north-west and the constant disputes between the numerous war- 
like clans, fragmented throughout the rugged valleys of southern and central Tibet, were 
quickly ended. The Sumpa (Supi), the Greater and Lesser Yang-t'ung (Upper and Lower 
Zhangzhung), the Bailan, the A-zha (T'u-yii-hun) and other peoples of Chiang stock were 
annexed in quick succession. The great reserve of energy of the Tibetan people, tempered 

7 Hsiian-tsang, 1985, Ch. 1, pp. 48, 54. See also Litvinsky (ed.), 1992. 



ISBN 978-92-3-103211-0 


The kingdom of Tibet 

FIG. 2. Kumtura. Bodhisattva. Musee Guimet, Paris. Photo: © R.M.N./© Jacques L'Hoir/Jean 

by an austere life in the wilderness, was set free and directed to politico-military expan- 
sion. This extended over the vast space of the Koko Nor-Tibetan plateau and lasted for 
some two and a half centuries. 

Early in the seventh century, gNam-ri-srong-brtsan (c. 570-620), btsan-po (chief) of 
the Yar-klung-spu-rgyal clan, had already laid the foundations of the new monarchy by 
bringing more than a dozen clans under his rule. 8 But the feudal customs of the tribal 
chieftains and the noble clans remained semiautonomous, although incorporated into the 
comparatively centralized state under the btsan-po. The death of gNam-ri-srong-brtsan 
provoked a rebellion. The brilliant new era in Tibetan history was ushered in with the 

8 There seem to have been 18 noble clans. A list of them is found in Bka' thang sde Inga, c. 1285, Vol. 5, 
fol. 7. See Tucci, 1949, pp. 737-8; Chang Kun, 1959-60, pp. 130-1, 153, note 10. 



ISBN 978-92-3-103211-0 Contents The kingdom of Tibet 

enthronement of Srong-brtsan sgam-po (c. 620-649 or 650), gNam-ri's son, as btsan-po. 
The new ruler quickly unmasked all intrigues and quelled all signs of open revolt. In 633 
Srong-brtsan sgam-po moved the royal residence from Lho-kha to Lhasa ('place of the 
god'), a step of great strategic importance for the rapid expansion that followed. 

According to Tibetan and Chinese records, Srong-brtsan sgam-po undertook sweeping 
legislative reforms with the assistance of his chief minister, sTongbrtsan yul-bzung of the 
mGar clan (Lu-dong-zang in Chinese literature, 1-661). sTong-brtsan served the btsan-po 
until the latter's death, and his office of Great Minister (blon-chen-bo) was filled by his 
descendants for decades. The btsan-po, fully aware of the presence of four powers in the 
monarchy - majesty (nMgathang); magic (dBu-ring; literally, the 'helmet' worn by the 
btsan-po at sacred functions); religious law (chos), which originated with the Bon-po; and 
political authority (chab-srid), which the ruler exercised through his officials 9 - set up new 
political institutions. Many of them were modelled on the T'ang administrative system and 
regulations: the ministers had, for example, credentials and emblems of rank. The Great 
Minister and the vice-Great Minister (blon-chenvog- ma) were followed by the 'inner' 
(nang-blon) and 'outer' (phyi-blon) ministers and a supreme judge (shal-ce-pa chen-po), 
each of whom had the insignia accorded to his position in the hierarchy. The 'inner' min- 
ister handled the internal affairs of the court while the 'outer' minister was in charge of 
relations between subordinate clans, external reconnaissance and the launching of punitive 
expeditions. The supreme judge oversaw the administration of justice; his position was 
similar to that of the 'minister of punishment' under the T'ang dynasty. In addition, there 
was an official (mngan-pon) who was responsible for the budget and a chief accountant 
(rtsis-pa-chen-po) in charge of book-keeping. Laws were drawn up, some of which were 
designed for the half-cultivator/halfherdsman activities of the population, as evidenced by 
the fragmentary Tibetan scrolls of legal documents found in Dunhuang. 10 

One of the most important reforms seems to have been the integration of the entire Tibetan 
territory, setting up 5 ru (dbu-ru, g 'yo-ru, g 'yas-ru, ru-lag and sum-pa-ru) and 61 stong-sde 
as unified military and administrative units. 1 1 This institution was supplemented by the oath- 
of-alliance, a characteristic practice among the Tibetans when subordinates pledged their 
loyalty to the nobles, and the nobles in their turn swore allegiance to the monarch. 

The king had subjects who were 'near to his heart' and others who were 'distant from 
his heart' . 12 They regularly exchanged oaths of fealty, accompanied by a ceremony of sac- 

9 Hoffman, 1975, p. 389. 

10 For example, 'Compensation Law on Hunters who Injure a Third Party', a Tibetan manuscript in the 
collection of the Bibliotheque Nationale of Paris, Fond Pelliot tibetain, PT 1,071. 

11 See Bacot et al., 1940^16; Uray, 1962, pp. 353-60. 

12 Stein, 1972, p. 132. 



ISBN 978-92-3-103211-0 Contents The kingdom of Tibet 

rifice, a minor one every year and a major one every three years. The words of the oaths, 
apart from expressing a feeling of awe for the deities called as witnesses, stipulated the 
rights and obligations of all the allied parties. This system - forming alliances by oaths and 
even concluding treaties - brought great advantages to the newly founded Tibetan regime. 
Such oath-taking rituals between the btsan-po and his ministers appear to have become a 
fixed procedure in the political life of Tibetan society. In addition, when the btsan-po pro- 
claimed some important policy, the oath-taking ceremonies were often used as the form for 
such proclamations. For example, Khri-srong lde-brtsan (755-797) held such ceremonies 
twice during his reign to make public the royal edict on advocating Buddhism. The cere- 
monies were also staged when establishing relations with the T'ang or with neighbouring 

Judging from the lists of names attending several large-scale alliance ceremonies, and 
found in the Tibetan chronicle discovered in Dunhuang, the system seems to have been 
instrumental in realizing the dynastic aspirations of Srong-brtsan sgam-po and his succes- 
sors. The essence of these alliances was their military strength: by joining the alliance, 
every clan inevitably became a military unit. The stong-sde and ru were in effect the 
designations for different clans under the command of the 'Supreme Commander of all 
Troops under Heaven'. 


In 634 the Tibetans established diplomatic relations with the T'ang court, sending envoys 
and tribute. Soon afterwards, Srong-brtsan sgam-po was granted a marriage with a Chinese 
princess, Weng-cheng Kung-chu, known to the Tibetans as Mun-chhang Kong-cho; she 
was sent to Tibet in 641 . Then followed a period of friendship lasting two decades between 
the Tibetans and the Chinese. Young people of the Tibetan nobility were sent to Ch'ang- 
an to study the Chinese classics, Chinese books were brought to Lhasa and there were 
constant attempts to translate and adapt Chinese classics and literary works into Tibetan. 
Some of these translated texts can be found in the Tibetan manuscript hoard of Dunhuang. 
At the same time other clans, including the Yangtong, Tiirgesh, Bolor and Nepal, also built 
up relations with Tibet by marriage. 

Within half a century the power of the Tibetan state had expanded westward towards the 
Pamirs, eastward towards Su-chuan, Yun-nan (Nan-chao) and northward to impinge upon 
the Koko Nor area and the Gansu corridor. After 663 Tibetan military incursions, especially 
those led by the ministers and commanders of the mGar clan (sometimes in alliance with 
the Western Turks), posed a direct threat to, and exerted pressure on, the Chinese western 
and southwestern frontier. With the deliberate aim of counterbalancing Tibetan influence 



ISBN 978-92-3-103211-0 Contents The kingdom of Tibet 

over the Western Regions, the T'ang court organized successive expeditions against the 
Tibetans both in the Gansu corridor and in the Tarim (670, 675-679, 692). In 692 the T'ang 
decided to station some 30,000 'protective' soldiers permanently at the Four Garrisons of 
the An-hsi Protectorate- General to defend the Tarim area and the communication route 
against the Tibetans in the south and their Turk ally in the north. 

In 710 another Chinese princess, Chin-ch'eng (Kyim-sheng Kong-cho), was granted 
to a Tibetan king. According to some sources, in 742 she gave birth to a son who was 
destined to become king: this was Khri-srong lde-brtsan. Under his reign, the Tibetans 
once again ruled over Gilgit (Drusha). The possession of Gilgit and its neighbour Baltistan 
was of great strategic importance: it allowed the Tibetans to control the main route from 
Kashgar through the Mintaka pass to Kashmir and the Indus valley, and made it possible 
for Tibet to establish direct contact with the Turkic tribes of the Tarim area and the Arabs 
of Central Asia. At the same time, tribute was paid twice by the Pala kings of Magadha 
and Bengal to Tibet, in 755 and in 756. According to some records, the Tibetans may also 
have invaded India in search of relics of the Buddha in Magadha and set up an iron column 
on the Ganges. 13 

In the reverse direction, the Tibetans took the opportunity of the An Lushan rebellion 
(755-763) to invade China proper and even captured Ch'angan, placing a boy-emperor on 
the T'ang throne for 15 days (in the eleventh month of 763). Then, availing themselves of 
the withdrawal of the Chinese garrison, they succeeded in occupying the Gansu corridor 
and the Four Garrisons of the An-hsi Protectorate-General (from 763 onwards). Thus the 
Tibetans attained their long-desired objective: complete control of the route through the 
Gansu corridor and the Tarim to Central Asia, Kashmir and northern India; and through 
modern Afghanistan to Transoxania and Iran. Their military presence in these vast areas 
also allowed them, from 791 onwards, steadfastly to oppose the challenge posed by the 
Arabs in Central Asia. 


Tibet's high-altitude environment, with a diversity of local conditions, gave its economic 
life a two-fold structure of cereal culture and animal husbandry, with people frequently 
alternating between the two modes of life according to the environment and the season. 
All political and cultural centres were located in regions of intensive cultivation. The rise 
of the first Tibetan monarchy was mainly due to its control of the most fertile cultivated 

13 Stein, 1972, p. 64. 



ISBN 978-92-3-103211-0 Contents The kingdom of Tibet 

regions, where highland barley, wheat, buckwheat and beans were grown. Two yoked oxen 
were used for ploughing, and irrigation and drainage techniques were known. 

Tibetan pastoral stockbreeding was more advanced than agriculture. Domesticated ani- 
mals comprised yaks, a yak-and-cow hybrid, goats, sheep, horses and a few pigs and dogs. 
The yak played (and continues to play) an important role - indeed, one could go so far as 
to say that without the yak, there would be no Tibetan culture. These animals roved around 
during spring and summer in search of pasture and water, and during autumn and winter 
they stayed on pastureland. Stories of levying 'taxes for oxen legs' reveal one aspect of the 
development of animal husbandry. 

Tibetan metallurgical techniques (in iron-working) had reached a very high level, sug- 
gesting a development over several centuries, and the skills of quench-hardening and grind- 
ing had already been mastered. Needless to say, these techniques ensured an ample supply 
of suits of armour and sharp swords for the troops. Even today, in their religious ritu- 
als, Tibetans still use large bells and elaborate golden utensils cast during the time of the 
ancient Tibetan monarchy. 

In the mid-seventh century, in response to an earnest request from Srongbrtsan sgam- 
po and his Chinese bride, Chinese craftsmen and artisans trained in the manufacture of 
rice alcohol, mill-stones, paper, ink and glass were sent to Tibet. Silk-worms and tea were 
also introduced. Simultaneously, Tibet fell under the influence of its western and southern 
neighbours: from India, for example, came astrological calculations and medical science. 


The most outstanding attainments during the reign of Srong-brtsan sgam-po were the cre- 
ation of a Tibetan script and the introduction of Buddhism. According to Tibetan tradition, 
Srong-brtsan sgam-po sent a young minister named Thon-mi-sambhota, sambhota of the 
Thon-mi clan, with other youngsters to the Kashmir area to study languages. After many 
vicissitudes, the minister succeeded in learning Sanskrit and several other languages of 
ancient India; he then made comparative studies of them before creating an alphabetic sys- 
tem for the writing of documents. Thon-mi-sambhota is also credited with compiling a 
Tibetan grammar on the Indian pattern. 


Before the introduction of Buddhism the religion of the Tibetans was Bon-po, which had 
many similarities with other primitive religions. Some scholars believe that Bon-po is a 
variant of Shamanism, while others insist on seeing the country of Zhang-zhung as the 
home of the Bon-po religion. Since most of Tibet's neighbours were Buddhist, the influ- 



ISBN 978-92-3-103211-0 Contents The kingdom of Tibet 

ence of Buddhism was strong. The introduction of Buddhism was marked especially by 
the occasion when Princess Weng-cheng of the T'ang dynasty and Princess Khri-brtsun, 
daughter of the king of Nepal, each brought a figure of the Buddha into Lhasa. Both 
princesses were married to Srong-brtsan sgam-po and propagated Buddhism among the 
Tibetans. After several generations, the aboriginal religions in the Tibetan area were either 
gradually displaced or became integrated into the more systematic and better-knit philo- 
sophical system of Buddhism. 

A decisive event in the history of Tibetan Buddhism was the adoption in 791 by King 
Khri-srong lde-brtsan of Indian Buddhism as the state religion. It was the culmination 
of a process in which Indian Buddhism replaced not only the Bon-po religion but also 
counteracted the influence in Tibet of the Chinese Ch'an tradition of Buddhism. The main 
task was the translation of Buddhist writings into Tibetan and the unification of Buddhist 
terms (the drafting of the Mahdvyut-patti in 814), an activity vigorously supported by the 
king. He was greatly helped by one Shang-shi, who brought Buddhist books to Tibet from 
China, and by gSal-snang, governor of a Tibetan province bordering Nepal who not only 
brought in Indian books but also persuaded the great Mahayana teacher Santiraksita to 
spend some time in Tibet. He ordained seven young men as 'the chosen ones' {sad-mi) to 
continue his work. On Santiraksita's departure, his place was taken by the towering figure 
of Padmasambhava from Uddiyana. During the same period, the great Buddhist temple of 
bSsam-yas was completed (755?). Important though the advance of Buddhism was, it did 
not go unhindered. Nor did it eliminate Bon-po beliefs, which were vigorously supported 
by some factions of the feudal nobility. 


Since 789 the Tibetan troops had been pushing towards Beshbalyk and were in fierce 
confrontation with the Uighurs. As the two sides plunged into an evenhanded war, their 
relations deteriorated irretrievably. The protracted warfare constantly forced the Tibetan 
regime to enlist Nanshao troops from the Yunnan to fight on the frontier. The Tibetans 
found themselves increasingly isolated, however, and the domestic situation reached a cri- 
sis. Although the army remained powerful, it could not avert the regime's final defeat. The 
palace coup of 846 revealed the serious corruption in the Tibetan ruling class, and frequent 
factional conflicts and religious strife sapped the morale of its troops and caused the col- 
lapse of the once powerful alliance of clans. Under the pressure of uprisings by the nobles 
and the common people in various areas, as well as in their own territory, Tibet's glorious 
early history of more than two centuries came to an end. 



ISBN 978-92-3-10321 1-0 Contents TOKHARISTAN AND GANDHARA 



J. Harmatta and B. A. Litvinsky 


Tokharistan 362 

Kapisa- Gandhara 365 

Zabulistan 371 

The fight for independence 372 


Ethnic groups and languages 376 

Writing systems and literature 379 

The provinces and their rule 381 

Coinage 382 

Cities, architecture and art 383 

See Map 7. 



ISBN 978-92-3-10321 1-0 Contents HISTORY OF THE REGIONS 

Part One 


{J. Harmatta) 

Trade, and above all the silk trade, played a major role in the economic life of the states 
of Central Asia in the sixth and seventh centuries. Political and military events, for both the 
sedentary and the nomadic peoples of the time, were largely determined by the struggle for 
control of the Silk Route. About the middle of the sixth century, the Hephthalite kingdom 
controlled (and derived considerable economic benefit from) the most important sections 
of the route, which led across Central Asia together with its branches from the Tarim basin 
to the Aral Sea in the west and to Barygaza-Broach in the south. 

At that very time, however, a powerful rival appeared in Central Asia, the Turk tribal 
confederation (see Chapter 14). The Turks first came to the Chinese frontier fortresses to 
barter their products for silk in c. 545 but they were refused. After their military victory 
over the T'ieh-le and the Juan-juan, however, they received great quantities of silk from 
the Chinese states. From 569 the Chou court supplied the Turks with 100,000 bales of 
silk a year. 1 As the Turks accumulated great stores of the precious material, their efforts to 
develop the silk trade and to gain control over the Silk Route became ever more aggressive. 
As a consequence of their economic interests, and in alliance with the Sasanians (who 
shared these interests in many respects), the Turks overthrew the Hephthalite kingdom, but 
could only take possession of the territory of Sogdiana. The Sasanians secured Chaganiyan, 
Sind, Bust, Rukhkhaj, Zabulistan, Tokharistan, Turistan and Balistan as vassal kingdoms 
and principalities. 2 

Thus, the Turks took possession of great sections of the Silk Route in Central Asia. 
In spite of their military success, the Turks and their Sogdian merchants could only sell 
their silk stocks to the Sasanians, who refused, however, to establish trading relations with 
them. At first, the Turks tried to establish trading relations with the Byzantines and to 

1 Ecsedy, 1968, pp. 131-80. 

2 Harmatta, 1969, p. 401 and note 71. 



ISBN 978-92-3-10321 1-0 Contents HISTORY OF THE REGIONS 

sell their silk stocks directly to them. But the steppe route - starting from Sogdiana and 
crossing the deserts to the north of the Caspian Sea and the Volga to reach Byzantine 
territory on the south-eastern shores of the Black Sea, and thence by ship to Constantinople 
- proved too difficult for the nomads. It was also unsafe because, from their fortresses on 
the limes Sasanicus in the Caucasus, the Sasanians controlled the land to the north up to 
the Kuban valley. Thus the Turks soon reverted to military force. In 569-570 they launched 
a great military expedition against Sasanian Iran in which they conquered the territory of 
the former Hephthalite kingdom belonging to Iran in the form of vassal kingdoms and 
principalities. (The Sasanians were powerless to resist because they were also engaged in 
war against Byzantium.) Although there is no source that gives the details of the war waged 
by the Turks against Iran in 569-570, it is clear from the phrase 'Turkun wa Kabulu' (The 
Turks and [the people of] Kabul), in a poem written between 575 and 580 by the Arab poet 
al- c Asha, that the Turk army was operating in the Kabul-Gandhara area in 570. 3 

Later historical events show that the successor principalities of the Hephthalite king- 
dom, formerly annexed to Iran, accepted Turk supremacy and became vassals of the West- 
ern Turk kaghan. Thus, the southern section of the Silk Route was opened to the Turks and 
the Sogdian silk merchants, who were able to transport their merchandise to the harbours 
on the western shores of India. The taking of the city of Bosporus by the Turk army in the 
Crimea in these years was also designed to ensure control of the steppe Silk Route up to 
the Black Sea. 

The former Hephthalite territories were probably not yet under permanent military 
occupation at this time. Since the Turk army consisted of tribal military forces, the per- 
manent garrisoning of troops would only have been possible through the transfer of entire 
tribal groups and their livestock, providing them with an economic base. Thus, the Heph- 
thalite principalities continued to exist as vassals of the Western Turk kaghans, while the 
Xingil dynasty ruled in Kabul and Gandhara. 

However, the Sasanians did not renounce their claim to eastern Iran nor did the Hep- 
hthalites abandon their aspirations for independence. According to the Pei-shih (Chap- 
ter XLIV, p. 4), both the Sasanians and the Hephthalites revolted against Tardu (Ta-t'ou) 
kaghan in 581 or 582. 4 Some years later, in 588-589, in a further war with the Heph- 
thalites, the Sasanian army, under their commander-in-chief Bahram Chobin, took Balkh 
and crossed the Amu Darya. In the battle against the Turk army coming to the aid of the 

i Harmatta, 1962, p. 133, note 5. 
4 Chavannes, 1903, p. 50. 



ISBN 978-92-3-10321 1-0 Contents HISTORY OF THE REGIONS 

Hephthalites, Bahram Chobin killed Ch'u-lo (*Cor), the Turk kaghan, with an arrow and 
obtained great booty. 5 

Bahram Chobin's military successes were to have no lasting consequences, however, 
because shortly after his victory he revolted against the Sasanian emperor, Hormizd IV 
(579-590). Nevertheless, Vistahm, who was appointed governor of Khurasan by Khusrau 
II (590-628) after Bahram Chobin had been defeated, compelled the Hephthalite rulers 
Shaug and Pariowk to acknowledge his supremacy. Later, in 595 or 596, however, Vistahm 
was treacherously murdered by Pariowk. 

The troubled years between 591 and 596 led to the Western Turk kaghans' decision 
to change the system of vassal Hephthalite principalities in eastern Iran and to submit the 
territory of the former Hephthalite kingdom to direct Turk rule. The realization of this plan 
was delayed, however, because of internecine wars between the Northern and the Western 
Turks. The accession of Jig (Shih-kuei) kaghan in 611 stabilized the internal situation of 
the Western Turk Empire. When war broke out between the Sasanians and the Hephthalites 
in 616-617, the Turk kaghan sent an army to the aid of the Hephthalites, won a great 
victory over the Sasanians and advanced as far as Ray and Isfahan. 6 

Two interesting material relics connected with the Turk invasion of Iran have recently 
become known. The Foroughi Collection of Sasanian seals includes a remarkable speci- 
men with Middle Persian and Turk runic inscriptions. The Middle Persian legend runs as 
follows: (1) zyk, (2) hhn, (3) GDH (Zig kaghan, glory!), while the Turkic text runs: (1) 
b(a)q (e)s eb, (2) qi'y (U)g (o) nkii (Take care for companions, house, settlement; make a 
good name for yourself!). This is clearly a seal of Jig kaghan, destined for the administra- 
tion of the conquered territories. The Middle Persian legend was probably prepared with 
the help of Sogdian scribes because the spelling hhn of the word 'kaghan' reflects Sogdian 
orthography (Sogdian y' y'n versus Middle Persian h'k'n). The runic text gives the norms 
of royal behaviour for the Turk kaghans, in concise form. The other noteworthy material 
relic of the Turk invasion of Iran is a medal representing Jig kaghan in profile with the leg- 
end: (1) GDH 'pzwn zyk, (2) MLK' 'n MLK' (Glory, growth! Zig King of Kings), 7 which 
was probably minted to commemorate his victory. 

It is clear from the inscriptions that the Western Turk kaghans intended to annex the 
eastern Iranian territories to their realm. In spite of their military success, however, they 
failed to realize their plans. For unknown reasons, the Turk army was recalled by Jig 

5 Marquart, 1901, p. 65; Markwart, 1938, pp. 138 et seq., 141 et seq., 153 et seq.; Czegledy, 1958, p. 24. 

6 Noldeke, Tabari, 1973, pp. 435, 478 et seq. 

7 It was published by Gobi, 1987, pp. 276 et seq., PL 39, Fig. 2, who could not, however, read the name 
of the king and erroneously dated the medal (anonymous in his opinion) from Islamic times. 



ISBN 978-92-3-103211-0 Contents Tokharistan 

kaghan. Thus Smbat Bagratuni, the Persian commander of Armenian origin, was again 
able to defeat the Hephthalites, killing their king in single combat. 

The definitive annexation of Tokharistan and Gandhara to the Western Turk Empire 
was to take place some years later, in c. 625, when Sasanian Iran became involved in the 
war against Byzantium that ultimately led to its eclipse. 8 The Western Turk army of T'ung 
Yabghukaghan advanced to the River Indus, took possession of the most important cities 
and replaced the Hephthalite dynasties with Turk rulers. This event was commemorated 
by a medal minted probably by Tardus/za<i, the new Turk ruler of Tokharistan, in honour 
of T'ung Yabghu kaghan, with the legend GDH 'pzwt' yyp MLK' 'n MLK' (The glory 
increased, jeb (= Yabghu) King of Kings!) 9 

Of the territories annexed in c. 625 by the Western Turk Empire, Khuttal and Kapisa- 
Gandhara were independent kingdoms after the disintegration of the Hephthalite kingdom. 
The Hephthalite kings bearing the title xingil of Kapisa-Gandhara continued the coinage 
of the Hephthalite kings of Tokharistan. The names of the kings Khingila II, Purvaditya, 
Triloka, Narana, Narendra I and Narendra II are attested by the legends of their coins. 
All the coin legends are written in the Brahml alphabet and all kings (with the exception 
of Khingila) bear Indian names. This is clear evidence of the slow Indianization of the 
Hephthalite royal dynasty during the sixth century. The same is true of the Hephthalite 
princes of Khuttal, who also minted coins with Indian legends: jayatu Baysara Khotalaka 
(Be victorious Baysara, [Lord] of Khuttal!), jayatu Baysara (Be victorious Baysara!) and 
sri Vasyara (His Highness Vasyara!). 10 

The last Hephthalite king of Kapisa-Gandhara, Narendra II, bears on his coin (Cabinet 
des Medailles 1974.443) a crown decorated with a bull's head. Since the bull's head also 
appears on the coins of the Turk yabghus of Tokharistan, this symbol clearly implies the 
recognition of Turk sovereignty. The appearance of the bull's head among the royal sym- 
bols of the Western Turk kaghans probably goes back to the title buqa (bull) adopted by 
Tardu kaghan after becoming the sole ruler of the entire Turk Empire in 599. 


Compared with Kapisa-Gandhara, Tokharistan (with its capital, Balkh) lost much of its 
former importance. Although the Hephthalite ruler of Balkh bore the Bactrian title sava 

8 For the connection of this war with the struggle for the Silk Route and the events in Central Asia, see 
Harmatta, 1974, pp. 95-106. 

9 Harmatta, 1982, pp. 167-80. 

10 Humbach, 1966, pp. 31, 58. 

11 Chavannes, 1903, p. 51. 



ISBN 978-92-3-103211-0 Contents Tokharistan 

(king), the name of his son, Pariowk (in Armenian, clerical error for *Parmowk) or Bar- 
muda, Parmuda (in Arabic and Persian, clerical error for *Barmuka, *Parmuka) which 
goes back to the Buddhist title pramukha, shows that he was the lord and head of the great 
Buddhist centre Naubahar at Balkh. His dignity and power were thus more of an ecclesi- 
astic than a secular nature. The famous Barmakid family of Islamic times were apparently 
the descendants of the Hephthalite pramukhas of the Naubahar at Balkh. 

After the Turk conquest, all the principalities of the former Hephthalite kingdom came 
under the rule of the Tiirkyabghu of Tokharistan residing in Qunduz. The Chinese encyclo- 
pedia the Chih-fu-yiian-kuei 12 lists the kingdoms subject to the Turk yabghu of Tokharis- 
tan: Hsieh-yii (Zabulistan), Chipin (Kapisa-Gandhara), Ku-t'u (Khuttal), Shih-han-na 
(Chaganiyan), Chiehsu (Shuman), Shih-ni (Shignan), I-ta (Badhghis), Hu-mi (Wakhan), 
Hushih-chien (Gozgan), Fan-yen (Bamiyan), Chiu-yiieh-to-chien (Kobadian) and Pu-t'o- 
shan (Badakhshan). The Chinese pilgrim Huei-ch'ao, who travelled in these lands between 
723 and 729, asserts that in Gandhara, Kapisa and Zabulistan the kings and military forces 
were T'u-chiieh ( Turks). 13 This evidence clearly shows the immigration of a Turkic popu- 
lation into these territories. The settlement of the Karluks is attested by the Chinese sources 
and the immigration of both the Karluks and the Kalach is shown by the Arabic and Persian 
sources. 14 

The first Turk ruler of Tokharistan and the subjugated petty kingdoms was Tardus/zad, 
the son of the Western Turk T'ung Yabghu kaghan. When Tardu was poisoned by his wife 
a few years later, he was succeeded by his son Ishbara yabghu, who, as first among the 
Turk rulers, began to mint coins. His coin effigy represents him bearing a crown decorated 
with two wings and a bull's head. The legend on one of his coins (Cabinet des Medailles 
1970/755) runs as follows: obverse: sb'lk' yyp MLK' (Isbara Jeb [= yabghu] sah); reverse: 
pncdh hwsp' ([minted in his] 15th [regnal year at] Khusp). If Ishbara yabghu ascended the 
throne in c. 630, the coin would have been minted in 645 at Khusp, a town in Kuhistan. 
Another issue 15 was struck in the 13th year of Ishbara at Herat (Hare) and a third one in his 
20th year at Shuburgan. This shows that Ishbara's reign lasted to 650 and that at least three 
mints (at Khusp, Herat and Shuburgan) were working in the western part of Tokharistan 
during this period. In c. 650, however, Western Turk power declined and its fragmented 
parts became, at least nominally, vassal kingdoms and principalities of the T'ang Empire. 

The first Turk yabghu (king) of Tokharistan, confirmed by the Chinese emperor, was 
Wu-shih-po of the A-shih-na dynasty. By this time, however (653), the Arab advance 

12 Chavannes, 1903, pp. 250 et seq. 

13 Fuchs, 1938, pp. 444, 447, 448. 

14 Minorsky, 1937, pp. 347 et seq.; Czegledy, 1984, p. 216. 

15 Gobi, 1967, Vol. 3, issues 265/2, 265/1. 



ISBN 978-92-3-103211-0 Contents Tokharistan 

towards Central Asia had already begun. In 652-653 al-Dahhak b. Qais (al-Ahnaf), the 
commander of the advance guard of Amir c Abdallah b. c Amir, took Merv-i rud, conquered 
the whole of Tokharistan and agreed with the inhabitants of Balkh on the terms of their 
capitulation. Under the rule of the Umayyad caliph c Ali (656-661), however, the Arabs 
were driven from eastern Iran (even from Nishapur) and the rule of Peroz III, son of Yazd- 
gird III, was reestablished by the yabghu of Tokharistan in Seistan. 16 Under the reign of 
the caliph Mu c awiya (661-680), Balkh and Kabul were retaken by c Abd al-Rahman b. 
Samura, but Arab rule did not last long. 

As a consequence of the Arab invasions, the power of the Turk yabghu of Tokharistan 
was considerably weakened. After Ishbarayabghu, relations with China also seem to have 
been interrupted because of the Tibetan conquest of the Tarim basin (see Chapter 15). It 
was not until 705, under P'an-tu-ni-li, the yabghu of Tokharistan, that another mission was 
sent to the Chinese court. By that time, the yabghu had moved to Badakhshan because his 
capital, Balkh, and the central territories of his kingdom were occupied by the Arabs. Thus, 
Shuburgan, Khusp and Herat (where the mints had worked for the yabghus of Tokharistan) 
were lost and their coinage ceased to exist at the beginning of the eighth century. 

Accordingly, the two issues known besides that of Ishbara yabghu can only be dated to 
the second half of the seventh century. One of them (Cabinet des Medailles 1965.1915), 
which bears the legend sym yyp MLK' (Sem Jeb [=yabghu] sah) on the obverse and hpt 
Ipwlg'n' ([minted in his] 7th [regnal year at] Shuburgan) on the reverse, is evidence that 
Shuburgan was still among the possessions of the yabghu of Tokharistan. Another speci- 
men of the same issue 17 indicates hwsp' (Khusp) as the minting place on the reverse. Sem 
yabghu may be identified with Wu-shih-po, the first Turk king of Tokharistan. The Chinese 
spelling (taking the Chinese character po as a clerical error for mu) and its North- Western 
T'ang form . u o-si(J3)-m(u y) may well reflect a foreign prototype *Asem ~ *Asim. Count- 
ing the reign of Sem Wu-shih-mu as starting in 653, his 7th regnal year would be 659-660, 
i.e. a year before the repeated Arab invasions under Mu c awiya. 

The third issue (Cabinet des Medailles 1970.749), 18 which bears the legend gwn spr' 
yyb MLK' (Giin Ispara Jeb [= yabghu] sah) on the obverse, does not indicate either the reg- 
nal years or the mint. This striking phenomenon can probably be explained by historical 
events, in the course of which (as mentioned previously), the yabghu of Tokharistan with- 
drew to Badakhshan, while his central territories and mints came under Arab rule. After 
Giin Ishpara yabghu, whose reign may be dated to the last decades of the seventh century, 

16 Marquart, 1901, pp. 67 et seq.; Harmatta, 1971, pp. 140 et seq.; Daffina, 1983, p. 133. 

17 Gobi, 1967, Vol. 3, issue 266. 

18 Ibid., Vol. 3, issues 267-71. 



ISBN 978-92-3-103211-0 Contents Kapisa- Gandhara 

the coinage of the yabghus of Tokharistan came to an end and the region lost its political 
and military importance. 

It appears that even the rule of the Turk A-shih-na dynasty ceased at that time. P'an-tu- 
ni-li was succeeded as yabghu of Tokharistan by Ti-she, king of Chaganiyan, in 719. The 
later yabghus are only mentioned by the Chinese sources on the occasion of their missions 
to the T'ang court: in 729 Ku-tu-lu Tun Ta-tu (Qutluy Ton Tardu) asked for aid against the 
Arabs; 20 years later, Shih-li-mang-kia-lo asked for and received military aid against the 
Tibetans; and in 758 Wu-na-to came personally to the T'ang court and took part in the fight 
against the rebel An Lu-shan. 19 

Kapisa- Gandhara 

As mentioned above, the Hephthalite kingdom of Kapisa-Gandhara managed to preserve 
its independence even after the annexation of the western territories of the Hephthalite 
kingdom, first by the Sasanians and subsequently by the Western Turks. At the time of the 
Western Turk conquest in c. 625, the last ruler of the Xingil dynasty in Gandhara, Narendra 
II, recognized the supremacy of T'ung yabghu kaghan and thus maintained his throne. 

According to the report of Hsiian-tsang, the Chinese Buddhist pilgrim, the royal dynasty 
of Gandhara was extinct by the time of his visit in 630 and the land had come under the 
rule of Kapisa. 20 From Hsiian-tsang's account it becomes clear that prior to his arrival, the 
authority of the Xingil dynasty had been confined to Gandhara, while in Kapisa another 
prince, probably of Western Turk origin, was ruling and only united the two kingdoms 
under his rule after the death of Narendra II. There is no other evidence for the separation 
of Kapisa from Gandhara prior to the Western Turk conquest. At the time that the Western 
Turks advanced to the Indus, in c. 625, Kapisa was probably separated from Gandhara and 
entrusted to a Western Turk prince who also became ruler of Gandhara after the extinction 
of the Xingil dynasty. 

According to the T'ang shu, the king of Kapisa and Gandhara in 658 was Ho-hsieh- 
chih, 21 whose name (North- Western T'ang xdi- yij-tsi < Turk *Qar yi'laci) clearly points 
to Turkic origin. At first, he may have been appointed king of Kapisa, then, after some years 
(but before 630) he succeeded Narendra II even in Gandhara, where his accession may have 
been facilitated by a marriage alliance. The new Turk dynasty adopted the Hephthalite 
royal title xingil and regarded themselves as the heirs to the Xingil dynasty. 22 

19 Chavannes, 1903, pp. 155-8; Chavannes, n.d., p. 95. 

20 Chavannes, 1903, p. 130. 

21 Ibid., p. 131 and note 4. 

22 Harmatta, 1969, pp. 404-5. 



ISBN 978-92-3-103211-0 Contents Kapisa- Gandhara 

In spite of the emphasis on continuity, the new Turk dynasty of Kapisa- Gandhara began 
to strike a new coin type 23 on which the king is represented with a crown similar to that 
of Ishbara yabghu but decorated with only one moon sickle instead of two; the Brahmi 
legend is replaced by a Pahlavi one, running as follows: nycky MLK' (King Niza/i/uk). The 
minting of this coin type lasted for almost a century (c. 630 - c. 720). The same effigy was 
maintained by the subsequent issues although minor modifications in the form of the crown 
and the ear-pendant can be observed and the legend gradually became deformed. 24 In view 
of the long period of minting and the fact that the first ruler of the Turk dynasty of Gand- 
hara bore the name *Qaryilaci, the legend nycky MLK' cannot represent a proper name; it 
can only be interpreted as a title or a dynastic name. The reading nycky MLK' 25 is firmly 
supported by the report of the Chih-fu-yiian-kuei according to which Na-sai, king of Ko- 
p'i-shih, sent a delegate to the Chinese court. 26 Without doubt Ko-p'i-shih (Ancient Chi- 
nese kd-b 'ji-sie) is the Chinese transcription of Kapisi (the kingdom of Kapisa-Gandhara) 
while Na-sai (Ancient Chinese nd-sok ) may well reflect the Bactrian variant *Nazuk of the 
name *Niziik. The reading nycky had previously been identified with the name of Tarkhan 
Nizak, the ruler of Badhghis. A thorough revision of the palaeographic and historical evi- 
dence, however, has revealed the true form of the latter to be Tirek, 27 a name of Tiirgesh 

When the supposed connection between nycky MLK' and Tarkhan Nizak is dropped, the 
relation of the Niziik dynasty with the tribal aristocracy of the Western Turk tribe A-hsi- 
chieh Ni-shu Szu-kin (*Askil Niziik Jigiri) becomes evident. The heads and nobles of this 
tribe bore the name Ni-shu (*Niziik, cf. Ni-shu Mo-ho shad, Ni-shu kaghan, Ni-shu Szu-kin, 
Ni-shu ch 'o). At the time of the Western Turk conquest, the royal powers and princely ranks 
in the successor states of the Hephthalite kingdom appear to have been distributed among 
the Western Turk tribal heads and nobles. Thus, the kingdom of Kapisa was entrusted to a 
member of the aristocracy of the Askil Niziik Jigin tribe. The element Niziik (going back 
to a Saka form *ntijsuka-, meaning 'fighter, warrior', from the Saka ndjs-, 'to fight') in the 
tribal name became the dynastic name of the kings of Kapisa-Gandhara, while their family 
name may have been Ho-hsieh-chih (*Qaryi'laci), which was borne by the first Western 
Tiirk yabghu of Kapisa-Gandhara. 

23 Gobi, 1967, Vol. 3, issues 198-205, 217-24. 

24 Gobi, 1967, Vol. 3, issues 198 (nycky MLK'), 200 iycky MLK'), 219 (cky MLK'). 

25 Besides, the readings npky and nypky are also possible, while nspky is impossible from a palaeographic 

26 Chavannes, n.d., p. 40. 

27 Esin, 1977, pp. 323 et seq. 



ISBN 978-92-3-103211-0 Contents Kapisa- Gandhara 

The Niziik dynasty of Kapisa-Gandhara separated into two branches in c. 670. Follow- 
ing a conflict between the king and his brother, the latter escaped to the Arab governor 
of Seistan, who permitted him to take up residence in the town of Zabul. The Arabs had 
already conquered Seistan in c. 650, but under the caliph c Ali both Seistan and Khurasan 
were lost. Under Mu c awiya, however, Seistan, Tokharistan and Kabul again came under 
Arab rule for a decade. After the death of the Arab governor c Abd al-Rahman b. Samura 
in 670 or 67 1 , the king of Kabul (Kapisa-Gandhara) expelled the Arabs from his territo- 
ries. At the same time, his brother (by now the ruler of Zabul) conquered Zabulistan and 
Rukhkhaj. Although he was then defeated by the new Arab governor al-Rabi c b. Ziyad, 
the Arab sub-governor of Seistan, Yazid b. Ziyad, later suffered a heavy defeat and fell in 
battle at Ganza (modern Ghazni). This clearly points to the strengthening of the kingdom 
of Zabul. Its ruler assumed the title zibil (earlier misreadings include zanbil, zunbil and rut- 
bil), going back to the ancient Hephthalite title zafiolo which was still borne by the kings 
of Zabul as late as the ninth century. 

The relationship between the two branches of the Western Turk Ho-hsieh-chih (*Qar 
y'ilaci) royal family, ruling in Kapisa and Zabul respectively, was far from peaceful. Accord- 
ing to the T'ang shu, Zabul (i.e. the branch of the family ruling in Zabul) extended its power 
over Kapisa-Gandhara after 711. This event is probably the basis of the legend concern- 
ing the origin of the Turk Shahi dynasty of Kabul, as told by al-Biruni in his India three 
centuries later. According to this legend, the founder of the dynasty ( Barhatakin) hid in 
a cavern and then unexpectedly appeared before the people as a miraculous being, thus 
coming to power. 

It is clear that the story of Barhatakin, with its cavern motif, represents a late echo of the 
legend of origin of the Turks (see Chapter 14 Part One) according to which their ancestors 
lived in a cavern. The real historical event, however, was quite different. According to the 
Chinese pilgrim Huei-ch'ao, who visited Gandhara between 723 and 729 (i.e. a decade 
after the event), when Wu-san T'e-chin Shai was ruling there: 

the father of the T'u-chiieh [Turk] king surrendered to the king of Chi-pin [Kapisa- Gand- 
hara] together with all sections of his people, with his soldiers and his horses. When the 
military force of the T'u-chueh strengthened later, he killed the king of Chi-pin and made 
himself lord of the country. 28 

Accordingly, power in both Zabul and Kapisa-Gandhara was concentrated in the hands of 
the same line of the Qaryi'laci royal family. Indeed, Huei-ch'ao explicitly states that the 

28 Fuchs, 1938, p. 445. Fuchs did not realize that at the time of Huei-ch'ao's visit, it was Wu-san T'e-chin 
Shai and not Barhatakin who was king in Kien-to-lo (Gandhara). For the origins of the Turk Shahi dynasty 
of Kabul, see Stein, Sir Aurel, 1893, pp. 1 et seq. 



ISBN 978-92-3-103211-0 Contents Kapisa- Gandhara 

king of Zabul was the nephew of the ruler of Kapisa- Gandhara. In spite of his legendary 
garb, the founder of the new dynasty, Barhatakin, must have been a real person. His name, 
Barha, is a hyperSanskritism for *Baha, going back to Turkic *Baya, while takin repre- 
sents the Turkic title tegin. The name Bay a is well attested among the Western Turks (in 
Chinese transcription Mo-ho; North-Western T'ang mbd y- yd Turk Bay a) and the title 
tegin (in Chinese transcription t'e-ch'in) was also widely used by them. Thus, the name 
*Baya tegin in the legendary tradition may in fact be authentic. 

It follows from Huei-ch'ao's report that Barhatakin had two sons: one who ruled after 
him in Kapisa-Gandhara, and another whose son became king of Zabul. According to 
the T'ang shu, the king of Kapisa-Gandhara between 719 and 739 was Wu-san T'e-ch'in 
Shai. It is clear from the historical context that he was the son of Barhatakin. The Chinese 
transcription (Ancient Chinese .uo-sdn d'ok-g'ion sai) reflects the Iranian title *Horsan 
tegin sahi. The Chinese form of the name follows the Chinese word order, however, and 
may be interpreted as 'Tegin sahi of Horsan' . The word Horsan may be the Hephthalite 
development of Xvarasan ( Khurasan) and the whole title obviously means 'Tegin, king of 

The coming to power of the new dynasty of Barhatakin was reflected in the coinage. 
The characteristic effigy of the Nizuk kings was replaced by a new royal portrait. The king 
bears a new crown decorated with three moon sickles, or tridents, which indicates a return 
to Hephthalite traditions 29 and is a clear declaration of independence from the Turk yabghu 
of Tokharistan. In the first issues, the meaningless remnants of the legend nycky MLK' were 
retained, 30 although they later came to have a purely decorative function. 31 At last, a new 
legend written in the Bactrian alphabet appears on the coins: a pio pavo (His Highness 
the King). 32 Seemingly, the name of the king does not appear in the legend. According to 
al-Biruni, Barhatakin assumed the title 'sahiya of Kabul' on coming to power. He could 
therefore be identified simply as 'the sahi' in the coin legend. Perhaps simultaneously, he 
also minted coins with the Brahml legend sri sahi (His Highness the sahi [King]). 33 

Barhatakin was followed by his son Tegin shah, who was ruler of Kapisa- Gandhara 
from 719. On his accession, Tegin assumed the high-ranking title Khurasan shah (king 
of Khurasan): this was a return to Hephthalite traditions because the two most important 
Hephthalite kings, Lakhana and Jabula, had both borne this title. Tegin shah continued the 
coinage of his father in so far as he retained the crown decorated with moon sickles to 

29 Gobi, 1967, Vol. l,p. 155. 

30 Ibid., Vol. 3, issues 225-31. 

31 Ibid., Vol. 3, issues 232-4. 

32 Ibid., Vol. 3, issues 236-7. 

33 Ibid., Vol. 3, issue 252. 



ISBN 978-92-3-103211-0 Contents Kapisa- Gandhara 

which he added two wings (the symbol of the farn, or royal splendour). He also took into 
account the various ethnic elements of his kingdom in the coin legends. His earliest issue 
(which can be dated to 721) has an exclusively Pahlavi legend to be read in the following 
way: obverse (10 h) GDH 'pzwt (2 h) 1. tkyn' bgy 2. hwtyp 3. hwl's'n MLK'; reverse (9 
h) TLYN (2 h) z'wlst'n (The royal splendour is increased! Tegin, the Majestic Lord, King 
of Khurasan, [minted in his] second [regnal year in] Zavulistan). The remarkable fact that 
the issue was minted in Zabulistan points to cordial relations between the kingdoms of 
Kapisa- Gandhara and Zabul. 

The next issue was again struck in Gandhara. Its legend is written partly in Bactrian, 
partly in Pahlavi and partly in Brahmi alphabets and in Bactrian (or Hephthalite), Middle 
Persian and Sanskrit languages. It runs as follows: obverse (2 h) a pi rayivo pavo; reverse 
(3 h) w'y (9 h) TLT', on both sides of the fire altar 1. srlla devil. Plnasri (His Highness 
Tegin, the King, [minted at] Way (hind) [in his] 3rd [regnal year]. The beautiful Queen 
Plnasri). The peculiarity of this issue lies in its mentioning the name of the queen. This 
was obviously a gesture towards the Indian population of Gandhara since the queen was 
of Indian origin and she enjoyed high status in Indian society. This is clearly shown by the 
Gilgit birch-bark manuscripts, which mention the name of the queen besides that of the 
king. Another interesting feature of this issue is that it was minted in Way (hind) (ancient 
Udabhandapura, and subsequently Hund), the capital of the sahi kings of Kabul. Its name 
occurs in its Middle Indian form here for the first time. 

The third coin type of Tegin shah was also minted in Gandhara. The style of the king's 
crown differs from those of the former issues. The legend is written exclusively in the 
Bactrian script and language: obverse (2 h) rayivo vcopaavo pavo; reverse (2 h) /povo 
v0S, (10 h) jxoppaoopo (Tegin, king of Horsano [Khurasan], [minted in] era-year 494, [at] 
Pursavur [Purushapura]). The development Pursavur of Purushapura almost exactly coin- 
cides with the medieval form of the name in Arabic literature, viz. Pur sowar and Pursor. 34 
No less noteworthy is the dating of the issue. The date is given in the Late Kushan era, 
which began in 23 1-232. 35 Accordingly, the issue was minted in 725-726. 

The fourth coin type of Tegin shah represents the king again bearing a new crown dec- 
orated with three moon sickles, two wings and an animal head. The legend on the obverse 
is written partly in Bactrian and partly in Brahmi; the reverse is in Pahlavi script. It runs as 
follows: obverse, in Bactrian: a gi pavo (His Highness the King); in Brahml: srihidibira 
kharalaca paramesvara srisahi tigina devdkdritam (His Highness, the hidibira, the Khar- 
alaca, the Supreme Lord, His Highness the sahi Tegin, the Majesty has [the coin] minted); 

34 Markwart, 1938, p. 109. 

35 Harmatta, 1969M, p. 425. 



ISBN 978-92-3-103211-0 Contents Kapisa- Gandhara 

reverse, inPahlavi: (a) tkyn' hwl's'nMLK' (b) hpf hpt't' (Tegin, king of Khurasan, [minted 
in the era-year] 77). 

This coin legend presents some interesting problems. As concerns the date, the year 
77 is apparently reckoned in the post-Yazdgird era, 36 which began in 652. Thus, the year 
77 corresponds to 729. Another question is raised by the word hidibira. This may be the 
same term as the Turkic title ilteber or elteber, which also has the variants iltber, ilber, 
attested among others by the Chinese transcription hsieh-li-fa (North-Western T'ang Xi(j)- 
Iji-pfvyj), reflecting a foreign prototype: * ilber. 31 The second problematic word of the coin 
legend is kharalaca, which must surely be a name or a title. It can be identified with the 
family name *Qaryilac'i of the dynasty if we assume a development r y > r and i > a 
(*Qaryi'lac'i> *Qaralaca) in it, which is abundantly attested in Old Turkic. 38 

This recognition may dispel the confusion in both the Chinese sources and the scholarly 
literature about the Chinese name Ko-ta-lo-chih or Ko-lo-ta-chih for Zabulistan. According 
to the T'ang shu, the original name of Zabulistan was Ts'ao-chii-cha (*Jaguda < Javula); 
between 656 and 660 it was named Ho-ta-lo-chih; and then the empress Wu changed this 
name to Hsiehyii (North-Western T'ang Zi-ivyj < *Zi f}il). It is a notorious mistake of 
early scholarly research that, on the basis of a superficial phonetic resemblance, the quoted 
Chinese spellings were identified with the name Rukhkhaj, used in Arabic geographic lit- 
erature to denote ancient Arachosia (Middle Persian Raxvad). However, this identification 
is impossible for several reasons. 

First, Ho-ta-lo-chih was only used officially by the T'ang court from 656 to 660. It 
is, therefore, impossible for it to have been used instead of the official name Hsieh-yii 
in a document written in the imperial chancellery in 718-719. Second, the Arabic form 
Rukhkhaj is not attested before the tenth century. It developed as a guttural assimilation 
from Middle Iranian Raxvad < *Raxvag, but simultaneously the original form Rakhwadh 
was still used by Ibn Rusta and Maqdisi as late as the tenth century. Moreover, the phonetic 
change g < j'had not yet taken place in Arabic in the seventh century, as is clearly proved by 
the Middle Persian transcriptions of Arabic names. Consequently, a form * Rukhkhaj could 
not have existed in Arabic in the seventh century when the Chinese name Ko-ta-lo-chih 
came into being. 

The Chinese initial ko/ho (Ancient Chinese kdt/ ydt) clearly points to a foreign initial 
*q. The North-Western T'ang form of the name was kdi-d 'd(j)-ld-tsi and counting on q 
- y < q - d dissimilation, or on the confusion of the sign hsia with ta (which are very 

36 Gobi, 1967, Vol. l,p. 144. 

37 The resemblance of the form hitivira to the Turkic title elteber had already been noted by Humbach, 
1966, p. 60. 

38 Cf., for example, qar yuy (sparrow hawk) > qaruy, bal'iq (fish) > balaq, Gabain, 1950, p. 49. 



ISBN 978-92-3-103211-0 Contents Zabulistan 

similar), one can regard the Chinese form Kdi-d'd(j)-ld-tsi or kdi- ya-ld-tsi as an exact 
transcription of the Turk royal family name Qaryi'laci', which became the name of the 
country. The case of Zabulistan shows the Chinese practice of naming a country after the 
name or title of its ruler. This may date back to nomadic usage and is attested up to the 
time of the Mongols. The first Chinese name for Zabulistan was Ts'ao-chii-cha < *Jaguda 
~ Jagula, the Hephthalite form of the royal title and name Javula. After the accession of 
the Qaryi'laci, the name of the country became *Ko-ho-lo-chih (* Qaryi'laci). Finally, after 
the separation of the two branches of the Qaryi'laci dynasty and the establishment of the 
Zibil kingdom at Ghazni, the Chinese named the country Hsieh-yii-kuo, or 'the country of 
the Hsieh-yii [*Zivil]'. This name was retained during the eighth century because all kings 
of Zabulistan bore the royal title zibil. Consequently, in the coin legend of Tegin shah, the 
terms hidibira Kharalaca (just as the Chinese phrases Ko-ta-lo-chih hsieh-li-fa and Ko-ta- 
lo-chih t'e-ch'in) do not mean 'the elteber of Arachosia' and 'the tegin of Arachosia', but 
simply indicate the family name (Qar yi'laci' ~ Qaralaca) and the titles {elteber and tegin 
respectively) of the kings. 

The characteristic features of the coinage of Tegin shah can be seen as reflecting the his- 
torical situation, the rich cultural tradition and the ethnic composition of Kapisa-Gandhara 
at that time. The coin legends are written in all the important languages (Bactrian, Middle 
Persian, Sanskrit) and scripts (Pahlavi, Bactrian, Brahmi) of the country and their contents 
refer equally to Persian, Hephthalite, Turkic and Indian traditions of royal ideology. The 
same syncretism is seen in Tegin shah 's dating of coins - in regnal years to stress his 
independence, in the Late Kushan era referring to local traditions, and in the post-Yazdgird 
era to indicate his distance from the Sasanian dynasty. 


As a contemporary of Tegin shah, his nephew Zibil ruled in Zabulistan from 720 to 
c. 738. His name was registered in the T'ang court in two different forms, Shih-yii and 
Shih-k'ii, but both spellings represent variants of the same title and name. Zibil Shih-yii 
(North-Western T'ang, *Zi-ivyf) reflects the form Zibil ~ Zivil, also attested by the Arabic 
sources, while Shih-k'ii (North-Western T'ang Zi-kivyf) represents a form *Zigil, being the 
Hephthalite development of Zivil. 

The independence, importance and power of Zabulistan are well illustrated by its coinage 
at that time. In this respect too, Zibil was independent from his uncle, Tegin shah. He cre- 
ated an effigy based on Sasanian traditions and on the coinage of the Arab governors, a 
phenomenon which reflects the fact that his interests lay towards the west, while his Indian 



ISBN 978-92-3-103211-0 Contents The fight for independence 

links are only represented by a short legend in Brahmi. The legend of his coins runs as fol- 
lows: obverse (1 h) yypwlhwtyp '(11 h) GDH (9 h) 'pzwt; on the rim around (1 h) PWN SM 
Yyzt' (3 h) yypwl bgyhwtyp' (6 h) whm'n'n mlt'n (9 h) MLK'\ reverse (11 h) snvakhudevah 
(1 h) pncdh z'wlst'n (3 h) 'pi plm'n yzd'n (King Jibul, [his] glory increased! In the name 
of god, Jibul, the Majestic Lord [is] King of brave men - His Highness the Majestic Lord 
- [minted in his] 15th [regnal year in] Zavulistan, by the order of the gods). 

Coin issues of Jibul ~ Zibil are so far known from his 2nd, 9th, 10th and 15th regnal 
years. It is very likely that he died shortly after his 15th regnal year (corresponding to 
735) because his son Ju-mo-fu-ta ascended the throne in 738. 39 In spite of the apparently 
entirely different form of his name in Chinese spelling, the new king of Zabulistan again 
bore the name or title Jibul. The North-Western T'ang form of Ju-mo-fu-ta was Ji-mbui 
pfvyi-d'di, which clearly reflects a foreign prototype *Jibul Pirdar (Elder Jibul), probably 
to be distinguished from a 'Junior Jibul'. 

The fight for independence 

One year later, in 739, Tegin shah abdicated the throne of Gandhara in favour of his son, 
Fu-lin-chi-p'o (also known as Fromo Kesaro, the Bactrian form of his name). 40 In this 
name, there is a confusion between the signp'o and so; accordingly, the correct form is 
Fu-lin-chi-so (North-Western T'ang pfvyr-humkie- sd) in which it is easy to recognize the 
Iranian name *From Kesar (emperor of Rome [=Byzantium]). This name implies an anti- 
Arab programme and propaganda at the time, which might be explained by Fromo Kesaro's 
having entered into manhood as an er at (meaning 'man's name') in 719, the year in which 
a Byzantine delegation travelled through Tokharistan on their way to the Chinese emperor 
and informed the kingdoms of Central Asia of the great victory they had won over the 
Arabs the previous year. 41 

The coinage of Fromo Kesaro (Fig. I) 42 is more closely connected with that of the 
Late Sasanians and of the Arab governors than with that of Tegin shah. The legends are 
written only in Bactrian and Pahlavi scripts and languages. They run as follows: obverse 
(1 1 h) (1) GDH (2) 'p < zwt > (2 h) (1) bg (2) hwtyp (The glory increased! The Majestic 
Sovereign); on the rim around, <ppo/xo xrjcrapo /3ayo xo^Srjo (Fromo Kesaro, the Majes- 
tic Sovereign); reverse (10 h) ST' (2 h) hwndy ([minted in his] 6th [regnal year at] Hund). 

39 Chavannes, 1903, p. 210 and note 1. As Chavannes noticed, the death of Shih-yu (*Zivil) and the acces- 
sion of Ju-mo-fu-ta could also have taken place two to three years earlier. 

40 Chavannes, 1903, p. 132. 

41 Harmatta, 1969, p. 412. 

42 Mochiri, 1987, PI. XXI, 125. 



ISBN 978-92-3-103211-0 


The fight for independence 

FIG. 1. Coin of Fromo Kesaro (obverse and reverse). (Courtesy of M. I. Mochiri.) 

This is the latest issue of Fromo Kesaro known so far to have been minted in Hund (ancient 
Udabhandapura) . 

The coinage of the kings of Zabulistan and Kapisa-Gandhara bears witness to the eco- 
nomic and political force and importance of both countries. They were able to preserve 
their ethnic and cultural identity and successfully fought for independence against the Arab 
conquerors. Arab rule was firmly established in Seistan, Badhghis, Gozgan, Tokharistan 
and Transoxania and even in Sind by the beginning of the eighth century. Nevertheless, and 
in spite of Qutaiba b. Muslim's tax-collecting expedition against Zabulistan in 710-711, 
both Zabulistan and Kapisa-Gandhara stood as islands in the sea of Arab predatory raids. 
It was only towards the end of the eighth century that both lands formally acknowledged 



ISBN 978-92-3-103211-0 Contents The fight for independence 

the supremacy of the Umayyad caliph al-Mahdi and the true conquest of Kabul did not 
take place until the end of the ninth century. 

An important recent discovery has provided a surprising insight into the events of this 
epoch. On the coins of some Arab governors, a Bactrian text overstruck on the rim has 
been discovered. 43 The reading of the text is as follows: (ppofio Kijaapo fiayo xoaSr/o 
klSo /So xaz iicavo /opyo o<5o crao /3o oa/3ayo aro i /xo /Jo yaivSo ( Fromo Kesaro, 
the Majestic Sovereign [is] who defeated the Arabs and laid a tax [on them]. Thus they sent 
it.). These coins formed part of the tax paid by the Arabs to Fromo Kesaro and were over- 
struck with a legend telling of his victory over them. Obviously, this event occurred during 
the reign of Fromo Kesaro (739-746) and may have contributed to his transformation in 
later historical tradition 44 into the Tibetan national hero Phrom Ge-sar, whose figure still 
survives today in the folklore of the territory of ancient Gandhara. 

The memory of the taxes paid by the Arabs has also been preserved in the Tibetan 
historical tradition according to which two Ta-zig (=Arab) kings, La-mer-mu and Hab- 
gdal, 'having taken kindly to Tibetan command, paid punctually without fail their gems 
and wealth' , 45 La-mer-mu may be an abridged form of the name c Amr b. Muslim, while 
Hab-gdal may have preserved the memory of c Abdallah b. al-Zubair. The latter evidence 
may also illustrate the successful resistance of the Gandharan population against the Arab 
conquest. However, the struggle was not decided here but in the far north at Talas, where 
the Arabs and Turks won a decisive victory over the Chinese army in 75 1 . 

Beside the most important successor states of the former Hephthalite kingdom (that 
is, Tokharistan, Kapisa-Gandhara and Zabulistan), some minor principalities also played 
a remarkable historical role during the time of the Arab conquest. Thus, Badhghis sur- 
rendered to the Arabs at an early date, but its energetic ruler, Tarkhan Tirek, continued the 
struggle until his death in 709. More successful was the resistance of Khuttal and Bamiyan, 
which disposed of greater military forces. The kings of Khuttal also struck coins, the land 
having had a tradition in this respect since the Late Hephthalite epoch. By the time of 
Huei-ch'ao's visit in the 720s, Khuttal already acknowledged Arab supremacy. 

To the north of Gandhara were two small states of great strategic importance: Great 
Po-lii and Little Po-lii according to the Chinese sources. The routes leading through these 
countries were equally significant for T'ang China and Tibet, and as a consequence of 
the Arab conquest of Khurasan, the arduous Silk Route connecting India directly with the 
Tarim basin became of vital importance. The Chinese name Polii (North-Western T'ang 

43 The discovery was made by Humbach, 1987, pp. 81 et seq. 

44 Harmatta, 1969, pp. 409 et seq. 

45 Thnmac 1 Q^S Vnl 1 r> 97^ 

naiiiiaua, iyvy, pp. tu? ci sc 

45 Thomas, 1935, Vol. 1, p. 273 



ISBN 978-92-3-103211-0 Contents The fight for independence 

Bui-lu) reflects the local form, Bolor (noticed later by al-Biruni), which goes back to the 
form *Bhauttapura (city of the Bhauttas), the latter being a Sanskrit term used for the 
Tibetans. 46 The population of the two Bolor (Po-lii) states consisted, however, of different 
ethnic elements: Tibetans, Dards and Burushaskis. It is interesting to note that the name 
Gilgit occurs in the Chinese sources for the first time during this epoch, appearing in the 
form Nieh-to in one text and Nieh-ho in another. Since no confusion of the sign to with 
ho seems possible, one sign is obviously missing from both spellings here. The correct 
form is therefore Niehho- to ( North-We stern T'anggz'i- y"d-td), which is a rather exact 
transcription (*Gilgat) of the name Gilgit. 

The conflicting Chinese and Tibetan interests led to China's military intervention in 
Gilgit in 747. Commanded by Kao Hsien-chih, a Chinese general of Korean origin, the 
Chinese forces won a decisive victory over the Tibetans and thus secured their routes to 
Khurasan and Gandhara. 47 

The period from 650 to 750 was a critical epoch in the history of Central Asia. The 
eclipse of Sasanian Iran, and the Western and Northern Turk empires, the crisis of the 
Byzantine Empire and the decline of T'ang China on the one hand, and the rise of the 
Arab caliphate and Tibet on the other hand, clearly indicate major historical changes. On 
the ruins of the ancient great empires, a new world was in the making. However, several 
centuries were to elapse before the emergence of significant new cultural achievements. 

46 Chavannes, 1903, pp. 149 et seq.; Markwart, 1938, pp. 103 et seq.; Fuchs, 1938, pp. 452 et seq.: (Khut- 
tal), p. 443 (Great Bolor), p. 444 (Little Bolor). 

47 Stein, Sir Aurel, 1923, pp. 173-7. 



ISBN 978-92-3-10321 1-0 Contents Ethnic groups and languages 

Part Two 


(8. A. Litvinsky) 

Ethnic groups and languages 

The kingdom of the Kabul Shahis was multiracial, inhabited by many different peoples. 
A considerable part of the population was composed of sedentary speakers of: (i) Middle 
and New East Iranian languages, Late Bactrian, and the New Iranian phase - the Afghan 
language; and (ii) West Iranian languages in the Middle Iranian and New Iranian phases 
- Tajik or Persian. Sanskrit and Prakrit were widespread. A large group of the population 
used Indo-Iranian Dardic languages as their mother tongues. Of the aboriginal languages 
of the east of the region, the linguistically isolated Burushaski should be mentioned. Of 
particular importance are the Turks (see Chapter 14), who brought their language from the 
depths of Central Asia. Information is given below about those ethnic groups and languages 
not discussed in previous chapters. 

The origins of the Tajiks and of their language lie in remotest antiquity. According to 
the eminent Iranologist Lazard: 

The language known as New Persian, which may usually be called at this period by the name 
of dan or parsT-i dan, can be classified linguistically as a continuation of Middle Persian, the 
official, religious and literary language of Sasanian Iran. . . 

New Persian belongs to the West Iranian group. In its phonetic and even its grammatical 
structure, New Persian had changed little from Middle Persian. Its vocabulary had changed, 
however, because New Persian drew heavily on the East Iranian languages, especially 
Sogdian, and also on the Turkic languages and Arabic. 48 Middle Persian was widespread in 
Khurasan and some parts of Middle Asia, partly promoted by the Manichaean 
movement. At the time of the Arab conquest, New Persian had already appeared in 

48 Lazard, 1971; 1975, pp. 595-7. 



ISBN 978-92-3-10321 1-0 Contents Ethnic groups and languages 

Tokharistan. According to Huei-ch'ao (writing in 726), the language of Khuttal - one of 
the most important domains of Tokharistan, located in the south of modern Tajikistan - 
was partly Tokharian, partly Turkic and partly indigenous. 49 

In connection with the events of the first third of the eighth century, the Arab historian 
al-Tabari relates that the inhabitants of Balkh used to sing in the New Persian (Tajik) lan- 
guage. It is quite possible, therefore, that a third ('indigenous', according to Huei-ch'ao) 
language was current in Tokharistan in addition to Tokharian and Turkic. If that is the 
case, Parsi-i Dari would appear to have been in use in Tokharistan as early as the sixth and 
seventh centuries. After the Arab conquest, the Dari language also spread to other parts 
of Middle Asia and Afghanistan. Much later it divided into separate Persian and Tajik 
branches, and a third branch is sometimes identified too - the Dari that is the contempo- 
rary New Persian language of Afghanistan. Some 30 million people speak these languages 
today. Like its close relatives Persian and Dari, Tajik has a rich history documented by lit- 
erary sources. The wealth of literary and scientific writings created in the Middle Ages in 
Parsi, the literary language that is common to both the Tajiks and the Persians, is a cultural 
asset of the peoples of Iran, Afghanistan and Tajikistan. 50 

The Tajiks emerged as a people in the ninth and tenth (or perhaps the tenth and eleventh) 
centuries, but it was not until the first third of the eleventh century that the term 'Tajik' 
began to be applied to them. That too was when Tajik (Persian) literature was founded, and 
its first great representatives lived and worked in Middle Asia. 

Although the origins of the Afghans lie in very ancient times, 51 the first mentions of the 
Afghan people appear only in the sixth and seventh centuries. The Brhat-samhitd (XVI, 
38 and XI, 61) speaks of the pahlava (Pahlavis), the svetahuna (White Huns or Heph- 
thalites), the avagdna (Afghans) and other peoples. On his return journey from India, the 
Chinese pilgrim Hsiian-tsang travelled from Varnu (possibly modern Wana) to Jaguda in 
Ghazni, crossing the land of A-p'o-k'ien, 52 a word derived from Avakdn or Avagdn, mean- 
ing Afghans. In Islamic sources, the first reliable mention of the Afghans is found in the 
Hudud al- c dlam, which says of a settlement on the borders of India and the Ghazni district 
that 'there are Afghans there too'. Mention is also made of a local ruler some of whose 
wives were Afghan women. 53 The Afghan language, or Pashto, is one of the East Iranian 
groups. Among its characteristics, it contains a stratum of Indian words and its phonetic 
system has been influenced by Indian phonetic systems, which is not the case of other 

49 Fuchs, 1938, p. 452. 

50 Oransky, 1988, p. 298. 

51 Morgenstierne, 1940; Grantovskiy, 1963. 

52 Hui-li, 1959, p. 188. 

53 Hudud al- c dlam, 1930, p. 16-a. 



ISBN 978-92-3-10321 1-0 Contents Ethnic groups and languages 

Iranian languages. There are approximately 23 million Pashto-speakers in Afghanistan and 
Pakistan today. 54 

The mountains in the east of modern Afghanistan and the north of modern Pakistan 
were settled by Dards. They were known to the ancient Greek authors, who used several 
distorted names for them: Derbioi, Durbaioi, Daidala, Dadikai and Derdaios. 55 In their 
descriptions of India, the Puranas speak of the Darada in the same breath as the inhabi- 
tants of Kashmir and Gandhara. They are repeatedly mentioned in the Rdmayana and the 
Saddharmasmrtyupasthana, together with the Odra (the Uddiyana). In Tibetan sources, the 
Darada are known as the Darta. 56 

There are two groups of languages that are now generally known as Dardic. The first 
are the languages of Nuristan (a region of Afghanistan): they form an 'individual branch of 
the Indo-Iranian family belonging neither to the Indo- Aryan, nor to the Iranian group'. The 
second group of languages (particularly the Dardic) are 'part of the Indo-Aryan [group], 
though far departed in their development from the latter'. The two groups, however, have 
much in common in their 'structural and material features [phonetical, grammatical and 
lexical]'. 57 The Nuristani languages include Kati, Waigali, Ashkun and Prasun (or Paruni) 
and are chiefly spoken in Nuristan. The Dardic languages proper include Dameli, which is 
the link between the Nuristani languages and the Central Dardic. According to one classi- 
fication, the Central Dardic languages comprise Pashai, Shumashti, Glangali, Kalarkalai, 
Gawar, Tirahi, Kalasha and Khowar. The Eastern Dardic group is divided into three sub- 
groups containing the Bashkarik, Torwali, Maiyan, Shina, Phalura and Kashmiri languages. 
In the early 1980s Dardic languages were spoken by 3.5 million people in Pakistan, India 
and Afghanistan, of whom 2.8 million spoke Kashmiri, some 165,000 spoke Khowar and 
some 120,000 spoke Pashai. The Nuristani languages were spoken by around 120,000 
people. 58 

Burushaski is a completely distinct language: it stands at the confluence of three great 
families - the Indo-European, the Sino-Tibetan and the Altaic - but belongs to none of 
them. Its speakers live in northern Pakistan, in the region of the Hunza and Vershikum 
rivers, and number around 40,000. The language's morphological structure is very rich and 
the verb has a particularly extensive system of accidence. Burushaski is one of the old- 
est tongues, but its place in the system of ancient and modern languages remains obscure. 
Although a literary tradition may well have existed in the early Middle Ages, when 

54 Morgenstierne, 1942; Gryunberg, 1987. 

55 Francfort, 1985, Vol. 1, pp. 397-8. 

56 Tucci, 1977, pp. 11-12. 

57 Edelman, 1983, pp. 14-15, 35-6. 

58 Morgenstierne, 1944; 1967; 1973; Fussman, 1972: Gryunberg, 1980; Edelman, 1983. 



ISBN 978-92-3-103211-0 Contents Writing systems and literature 

Buddhism was widespread, no literary records have been found, which hampers attempts 
to reconstruct the language's past. There have been repeated attempts to trace its affilia- 
tions, and links with the Caucasian, Dravidian, Munda, Basque and other languages have 
been suggested, but from the standpoint of contemporary linguistics the case is not conclu- 
sive. Burushaski was unquestionably more current in ancient times and occupied a number 
of regions where Dardic languages are now spoken and where Burushaski acted as a sub- 
stratal or adstratal foundation. Grierson has even postulated that speakers of Burushaski or 
related languages once inhabited all or almost all the lands now held by Dardic-speaking 
tribes. 59 

Writing systems and literature 

We have considerable information about the literature and writing systems of the period. 
Hsiian-tsang reports of the writing system of Tokharistan: 

In the composition of its language [Tokharistan] differs somewhat from the remaining 
realms. The number of letters in its script is 25, they combine to form various combinations 
and with their help all may be reproduced. The script is read horizontally, from left to right. 
Literary works are composed in great quantity and exceed the Sogdian in volume. 60 

This refers to the Late Bactrian writing system (for its development and writing, see 
Chapter 6), which persisted in some parts of Tokharistan as late as the twelfth century. With 
time, changes obviously occurred in the Bactrian language and its various written records 
may reflect different dialects. 61 The script became increasingly cursive, some characters 
were identical in shape and some had several meanings (this is particularly true of the 
ligatures), making the script difficult to decipher. 

Among the more famous written records of Late Bactrian (sometimes called Heph- 
thalite) writing, mention should be made of two cursive inscriptions carved on rocks in 
Uruzgan (north-west of Kandahar in Afghanistan). According to Bivar, who published 
them, one speaks of a king of Zabul called Mihira(kula) and dates from around 500, 62 
although other scholars (Henning and Livshits) suggest a far later date in the eighth or 
ninth century. The Bactrian inscriptions in the Toe hi valley of north-western Pakistan are 
very badly preserved. The Tochi valley also has Arabic and Sanskrit inscriptions from the 
first half of the ninth century. The text of the Bactrian inscription, which is very cursive, 

59 Grierson, 1919; Zarubin, 1927; Lorimer, 1935, Vol. 1; 1938, Vol. 2; Klimov and Edelman, 1970. 

60 Pelliot, 1934, p. 50. 

61 Gershevitch, 1985, p. 113. 

62 Bivar, 1954. 



ISBN 978-92-3-103211-0 Contents Writing systems and literature 

cannot be read with confidence: Humbach's proposed reading is completely rejected by 
other scholars. 63 

Inscriptions have also been found on sherds and walls in Middle Asia (at Afrasiab, 
Zang-tepe and Kafyr-kala among others). Hsiian-tsang's account suggests that many more 
manuscripts existed than have yet been discovered. Nevertheless some have been preserved 
in East Turkestan, in the Turfan oasis. 

Brahml manuscripts are known from Sir Aurel Stein's discovery of the Gilgit birch- 
bark manuscripts, which were immured in a stupa some time between the fifth and the 
seventh century. They include a Pratimoksa-sutra, a Prajnaparamita and others. A mathe- 
matical manuscript found near Peshawar, the Bakhshali manuscript (see below), may date 
from the end of this period. 64 Other birch-bark manuscripts have been found in Zang-tepe, 
30 km north of Termez, where fragments of at least 12 manuscripts have been found. One 
of them bears a Buddhist text from the Vinaya-vibhanga. A fragment of birchbark man- 
uscript bearing a text of apparently Buddhist content has been found at Kafyr-kala in the 
Vakhsh valley. Mention should also be made of the Buddhist birch-bark manuscripts found 
at Merv and nearby at Bairam-Ali. The latter find consists of 150 sheets, both sides of 
which bear a synopsis of various Buddhist works, written in Indian ink. It was compiled 
for his own use by a Buddhist priest of the Sarvastivada school. 65 Sanskrit manuscripts of 
varied content, including medical materials, and dating from different periods have been 
found in the Bamiyan valley (see also Chapter 18). 66 

It was during the late eighth and early ninth centuries that the Sarada script was devel- 
oped on the basis of Brahml. In Afghanistan, two marble sculptures have been found with 
inscriptions which 'represent transition scripts from Brahml to Sarada' 67 and which date 
from the eighth century. The origin and chronology of the 'proto-Sarada script [are] far 
from being certain and [are] still open to speculation'. 68 In this regard, some materials 
from Bamiyan are of interest. 

The Bakhshali manuscript is written in Sarada script and was copied by five scribes, the 
chief of whom was Ganakaraja. It appears to have been a commentary on an earlier mathe- 
matical work and contains rules and techniques for solving problems, chiefly in arithmetic 
but also in geometry and algebra. The standard of knowledge in this field is indicated by 
the fact that the work treats square roots, geometric and arithmetic progressions and so on. 

63 Humbach, 1966, pp. 110-17; see Gershevitch, 1985, p. 93; Harmatta, 1969, p. 345. 

64 Kaye, 1927; Gilgit Buddhist Manuscripts, 1959-60, Parts 1-2; and others. 

65 Vorobyova-Desyatovskaya, 1983, pp. 63-8. 

66 Levi, 1932;Pauly, 1967. 

67 The Archaeology of Afghanistan, 1978, p. 244. 

68 Sander, 1989, pp. 108-12. 



ISBN 978-92-3-10321 1-0 Contents The provinces and their rule 

Grammars are also known. 'The oldest work of this school of grammar known to us is 
by Durga Simha who flourished in about 800 a.d. and has written a commentary entitled 
Durgavritti and a Tikd of it.' 69 

The provinces and their rule 

According to Hsiian-tsang, in the year 629 Tokharistan (Tou-ho-lo) measured approxi- 
mately 1,000 li from south to north and some 3,000 li from east to west. He reports: 

For many centuries past the royal race has been extinct. The several chieftains have by force 
depended for the security of their possessions upon the natural divisions of the country, and 
each held their own independently, only relying upon the natural divisions of the country. 
Thus they have constituted twenty-seven states divided by natural boundaries, yet as a whole 
dependent on the T'u-chiieh tribes [Turks]. 70 

Later reports paint a somewhat different picture. From the year 718 we have another 
Chinese report (see page 37 1 above). The yabghu's younger brother ruled over Po-lii (prob- 
ably Baltistan but possibly Gilgit). The capital of the 'dominion of the yabghu of Tou-ho-lo 
[Tokharistan]' was in the vicinity of modern Qunduz. 71 T'ang chronicles report that the 
state of Tokharistan had a 'select host of 100,000, all expert in battle'. 72 In Khuttal alone, 
there were reportedly 50,000 troops. 73 The rulers (muluk, pi. of mdlik, in Arabic sources) 
of some provinces bore specific titles. In the state of Uddiyana (valley of Swat), 'by cus- 
tom people are not killed. Serious crimes are punished by exile, while trivial offences are 
pardoned. There are no tributes or taxes.' 74 There were reportedly 5 cities in this state 
and the ruler lived in the city of Chu-meng-yeh-li. 75 Use was made of trial by ordeal. The 
ruler took decisions only after consulting the priests. 76 In 745 the ruler of Kapisa was also 
the ruler of Uddiyana. 77 Earlier, in 726, a kinsman of the ruler of Kapisa was the ruler of 
Zabulistan. 78 Earlier still, in the time of Hsiian-tsang, 10 provinces were under his rule. 79 
Thus, in the seventh century, Kapisa was a very powerful state. 

In the state of Bamiyan, 'the literature, customary rules and money used in commerce 
are the same as those of the Tukhara country [ Tokharistan]. Their language is a little 

69 Pandey, 1973, p. 240. 

70 Beat, 1969, pp. 37-8. 

71 Enoki, 1977, p. 88. 

72 Malyavkin, 1989, p. 68. 

73 Chavannes, 1903, p. 200. 

74 Malyavkin, 1989, p. 70. 

75 Ibid., p. 245. 

76 Bichurin, 1950, Vol. 2, p. 270; Chavannes, 1903, pp. 128-9. 

77 Enoki, 1977, p. 91. 

78 Fuchs, 1938, p. 448. 

79 Hui-li, 1959, p. 55. 



ISBN 978-92-3-103211-0 Contents Coinage 

different.' 80 The ruler of Bamiyan had a large and powerful army 81 and bore the title 'sher- 
i Bamiyan', while the ruler of Kabul province bore that of ratbil shah. 82 The capital of the 
state, or so al-Biruni bluntly asserts, was Kabul. Against this must be set the account of the 
Chinese Buddhist pilgrim Wu-k'ung, who visited these parts in the 750s and reported that 
'KapisT country had its eastern capital in Gandhara. [The] king resided in winter here and 
in summer in KapisT.' 83 


The coinage not only differed considerably from region to region, but was different in 
each of the provinces of Tokharistan. In what is now southern Tajikistan three variations 
of cast copper coins with central holes circulated: (i) coins of Tokharistan with legends in 
late cursive Bactrian (Hephthalite) script; (ii) coins with Sogdian legends; and (iii) coins 
without legends. Particularly noteworthy are the local imitations of Peroz drachms, some 
countermarked with Sogdian legends, which remained current as late as the mid-eighth 
century. 84 

In the part of northern Tokharistan that is now the Surkhan Darya region of Uzbekistan, 
different varieties of coins circulated. In Chaganiyan, silver coins of the Sasanian shahan- 
shah Khusrau I (53 1-579) were common because Khusrau's conquests had extended to this 
region. Subsequently, imitations began to be struck. Interestingly, both genuine coins and 
imitations were countermarked, some with a cursive Bactrian legend of the ruler's name, 
others with a miniature portrait and others again with a symbol (tamgha). Sometimes the 
same coin was countermarked several times, with one impression on top of the other. Later, 
copper coins of the local Chaghdn khuddt dynasty began to be issued. On the obverse was 
a portrait copying Khusrau I, in the margin three portraits of the Chaghdn khuddt and on 
the reverse a fire altar. On some coins the obverse bore a Bactrian legend; sometimes it 
merely carried the title khidev (ruler) or 'Khnar (or Enar) the khidev' . There were also cop- 
per coins bearing the likeness of the ruler and his consort. These are the characteristic coins 
of the Sogdian and Turkic states. Unlike similar coins from Chach (modern Tashkent) and 
Sogdiana, they bore a non-Sogdian inscription and another symbol. 

80 Beat, 1969, p. 50. 

81 Fuchs, 1938, p. 448. 

82 There is also a view that 'ratbil is the result of the corrupt scribe of the word Zabul' (Pandey, 1973, pp. 
73-4). In the edition of the TarTkh-i Sistan, the editor reports that the manuscript gives the word ZNBYL, 
supporting the reading Zunbll. See also Ibn Khordadbeh, 1889, p. 39; Kohzad, 1950. 

83 Levi and Chavannes, 1895, pp. 349-57. 

84 Davidovich and Zeimal, 1980, p. 74. 



ISBN 978-92-3-103211-0 Contents Cities, architecture and art 

In Termez, copper coins were struck bearing a portrait of the ruler on the obverse, and a 
symbol of a different shape from that used in Chaganiyan on the reverse. This coinage was 
probably issued by the local dynasty of Termezs/za/js. 85 

Although the coinage of Afghanistan and Pakistan has not been studied in such detail, 
issues of Vrahitigin (or Vahitigina) should be noted. These were silver coins (probably 
struck in the late seventh century) bearing the bust of the ruler and inscriptions in Bac- 
trian and proto-Sarada, the meaning of which was: 'Caused to be made by Sri Hitivira 
Kharalava, the Supreme Lord Sri Vahitigina the God'. On the reverse is a divinity crowned 
with a flame and a Pahlavi inscription. The ruler's crown comprises a wolf's head, indi- 
cating Turkic affiliations, while the divinity replicates the images on coinage of Khusrau 
II (590-628). Coins of this kind are found in the Indus valley, in northern Pakistan and 
in Afghanistan, including Kabul. Humbach 86 has suggested that Vahitigina is the same as 
Barhatakin, the founder of the Kabul Turk dynasty, of which al-Biruni reports, 'The Hin- 
dus had kings residing in Kabul, Turks who were said to be of Tibetan origin.' Sachau 87 
suggested that this name derived from the Hindu Brhatkina or Brhatketu (for linguistics, 
see pages 375-6 above). 

Cities, architecture and art 

The capital of the state of Kapisa-Gandhara (possibly, its winter capital) was Udabhanda- 
pura, now the settlement of Hund, situated on the right bank of the Kabul river. Most of the 
city was surrounded by a defensive rampart. Later, in the Islamic period, it formed a square 
and its total length measured 1.3 km. Each side had a central gate fortified with bastions. 
Traces of older fortifications have been discovered and there is also a well-preserved sec- 
tion of the old wall some 20 m long. Around the fortified portion, the remains of buildings 
have been found, indicating the great extent of the town. 88 

Although Balkh remained the capital of Tokharistan, there were many other large towns 
that acted as provincial centres. One of them, the Vakhsh valley centre now known as 
Kafyr-kala, has already been described (see Chapter 6). In this period, the city was charac- 
terized by a radical restructuring of the palace and residential quarters. 

Individual structures, including palaces (Kafyr-kala), castles (Balalyktepe, Zang-tepe, 
etc.), houses (Kala-i Kafirnigan) and, of course, Buddhist buildings, have been studied in 
considerable detail. Here we shall concentrate on Ajina-tepe (Fig. 2). This fully excavated 

85 Rtveladze, 1987, pp. 120-9. 

86 Humbach, 1966. 

87 Sachau, 1888. 

88 Another identification is possible: see Caroe, 1962, pp. 97-8. 



ISBN 978-92-3-103211-0 Contents Cities, architecture and art 

Buddhist monastery consists of two halves that made up a single complex of religious and 
residential buildings, each half occupying an area of 50 x 100 m. The south-eastern half, 
which formed the monastery proper, consists of a quadrangle of buildings around a square 
courtyard. In the centre of each side is an aiwdn (hall) and behind it a cella. The cella on the 
south-eastern side contained sculptures, including a 4-m-high statue of the Buddha, placed 
on figured pedestals. The other cellas were large halls, which served both as assembly 
rooms for the sangha (monastic community) and as refectories. The aiwdn?, were linked 
by winding, vaulted corridors from which passages led off into tiny cells. Some or all of 
the complex was two-storied. 

The second part could be called the temple. Its overall layout was identical, but there 
were no cells for the monks. In the central shrine there was a vast quantity of Buddhist 
sculptures on pedestals, or on the floor between. In each wall of the long, winding corridors 
there were three or four deep-set niches (Fig. 3), in which large statues of the Buddha sat 
in varied poses. At the end of the final corridor was a gigantic pedestal taking up almost an 
entire section, on which was a 12-m-high statue of a recumbent Buddha in Nirvana (Fig. 4). 
The vaulted ceilings of the corridors, and their walls, were covered in paintings and there 
were also paintings in the shrines (Fig. 5). 

The entire centre of the courtyard was occupied by the main stupa, which was star- 
shaped in plan and accessed by four staircases, one in the centre of each side. In the corners 
of the yard were miniature stupas of the same type, some ornamented with reliefs depict- 
ing small human figures (Figs. 6-10). 89 Buddhist temples have also been found in Kala-i 
Kafirnigan (where some excellent paintings and sculptures have been preserved) and in 
the palace complex at Kafyr-kala. Overall, there are grounds for speaking of a Tokharistan 
school of art, related to, but not identical with, the art of central Afghanistan. 90 

Bamiyan has already been described in Chapter 6. Here we shall say a few words about 
the Fundukistan complex, which has been ascribed to the seventh century. 91 The part that 
has been excavated includes a shrine and, linked to it by a vaulted passageway, another 
area consisting of several monastic cells, an assembly hall and other communal rooms. 
The shrine is in the form of a square hall with three deep vaulted niches along each side: 
it appears that there were originally just two on the entrance side. Between the niches are 
pilasters with Corinthian-style capitals. In the centre of the shrine there was a slender stupa 
with an arcade on each side of its pedestal. The building material consisted of large- sized 
blocks of pakhsa. Clay statues stood in the niches, whose surface was lined with murals. 

89 Litvinsky and Zeimal, 1971. 

90 Litvinsky, 1981. 

91 Carl and Hackin, 1959. 



ISBN 978-92-3-103211-0 


Cities, architecture and art 

FIG. 2. Ajina-tepe. Reconstruction of the south-eastern part of the complex. 

FIG. 3. Ajina-tepe. Axonometric projection of locations XXIV and XXV. Reconstruction. 



ISBN 978-92-3-103211-0 


Cities, architecture and art 

FIG. 4. Ajina-tepe. Hand of the 12-m statue of the Buddha in Nirvana. (Photo: © Vladimir Terebenin.) 

FIG. 5. Ajina-tepe. Mural painting. (Photo: © Vladimir Terebenin.) 



ISBN 978-92-3-103211-0 


Cities, architecture and art 

FIG. 6. Ajina-tepe. Torso of a Bodhisattva. Painted clay. (Photo: © Vladimir Terebenin.) 

The art of Fundukistan is characterized by vivid colours, bold foreshortening and elegance: 
although it betrays a powerful Indian influence, there is also a certain similarity with the 
art of Ajina-tepe and Kala-i Kafirnigan (Figs. 11 and 12). 

Buildings of the late period at Tepe Sardar, near Ghazni, are of similar date. In this 
large Buddhist monastery complex, the main stupa is surrounded by many miniature stupas 
and shrines, ornamented with clay bas-reliefs. There were several colossal statues of the 
Buddha, including one seated and one of the Buddha in Nirvana. In one shrine, which is 
in the Hindu style, a clay sculpture of Mahishasuramardini (a form of the Hindu goddess 
Durga) was found. Thus a Hindu element was inserted within the Buddhist context. It is 



ISBN 978-92-3-103211-0 Contents Cities, architecture and art 

FIG. 7. Ajina-tepe. Head of a Buddha. Painted clay. (Photo: © Vladimir Terebenin.) 



ISBN 978-92-3-103211-0 


Cities, architecture and art 

FIG. 8. Ajina-tepe. Head of a brahman. Painted clay. (Photo: © Vladimir Terebenin.) 



ISBN 978-92-3-103211-0 


Cities, architecture and art 

FIG. 9. Ajina-tepe. Head of a noblewoman. Painted clay. (Photo: © Vladimir Terebenin.) 

thought that this shrine is linked with the upper classes of society. 92 The remains of a Hindu 
shrine have also been found in Chigha Saray (or Chaghan Sarai) in the Kunar valley, dating 
from the eighth or ninth century. 93 

Hindu art is also represented by finds of marble sculpture such as a Shiva and Parvati 
(Umamaheshvara) from Tepe Skandar 30: 'It is carved from one block of white marble and 
represents the four-armed, three-eyed Shiva seated on Nandi, flanked by his consort Parvati 

92 Taddei, 1972; 1973; 1974. 

93 Van Lohuizen, 1959. 



ISBN 978-92-3-103211-0 


Cities, architecture and art 

FIG. 10. Ajina-tepe. Head of a monk. Painted clay. (Photo: © Vladimir Terebenin.) 



ISBN 978-92-3-103211-0 


Cities, architecture and art 

FIG. 11. Fundukistan. Two naga kings (seventh century). Musee Guimet, Paris. (Photo: © 
UNESCO/Lore Hammerschmid.) 

and Skanda standing at the left side of his mother.' 94 The group stands on a pedestal with 
two steps. On the upper step there is a three-line inscription in a transitional script between 
BrahmT and Sarada. It cites Shiva as Maheshvara. 95 Another fine example of Hindu art is a 
marble statue of Surya from Khair Khanah: 

The piece can be divided into upper, middle and lower parts. In the centre of the upper part is 
Surya, flanked by Danda and Pingala. In the middle part is the driver Aruna holding the reins 
of two horses whose backs are shown as they veer upwards to the right and left. The lower 
part is the pedestal. 96 

A whole series of other marble Hindu sculptures dating from this period has been 
discovered. 97 Taken together, they indicate a powerful Indian influence and the spread of 
non-Buddhist Indian religions. 98 

94 Kuwayama, 1976. 

95 Ibid., pp. 381-3. 

96 Hackin and Carl, 1936; Kuwayama, 1976, pp. 375-6. 

97 For the latest analytical review, see Kuwayama, 1976, pp. 375-407. 

98 The Archaeology of Afghanistan, 1978, pp. 291-2. 



ISBN 978-92-3-103211-0 


Cities, architecture and art 

FIG. 12. Fundukistan. Dressed Buddha (seventh century). 
Musee Guimet, Paris. (Photo: © UNESCO/Lore Hammerschmid.) 



ISBN 978-92-3-103211-0 Contents ZOROASTRIANISM 




Ph. Gignoux and B. A. Litvinsky 




Part One 


{Ph. Gignoux) 

Nothing would be known about Zoroastrianism ( Mazdaism) under the Sasanians had 
not one of its most outstanding religious leaders, the mobad (high priest) Kartir (Kirder), 
left four inscriptions carved in rock at various places in Fars: near Persepolis, at Naqsh-i 
Rustam (KKZ and KNRm) and Naqsh-i Rajab (KNRb), and to the south of Kazerun at Sar 
Mashhad (KSM) (Figs. 1 and 2). In them, the magus begins by describing his own career. A 
mere ehrpat (theologian) under Shapur I (241-272), he was appointed by Shapur's succes- 
sor, Hormizd I (272-273), as 'Magupat of Ohrmazd [ Ahura Mazda]' (a title that refers to 

See Map 1 . 



ISBN 978-92-3-103211-0 Contents ZOROASTRIANISM 

the supreme god, not the king himself), and eventually by Bahram II (276-293) as 'Magu- 
pat of the Blessed Bahram [i.e. referring back to Bahram I (272-276)] and of Ohrmazd'. 1 
It was under Bahram II that Kartir obtained his most important offices, including judicial 
appointments. He benefited from the great favour of the king, who not only authorized him 
to have inscriptions carved and decorated with his bust but even had him included among 
the royal family and court officials on several reliefs. 

The work of the magus was essentially, as he tells us in the inscriptions, to encourage 
the foundation of fire temples and increase the number of their attendants, and to combat 
religions other than Mazdaism: Manichaeism (see Part Two of the present chapter), which 
was beginning to spread under the vigorous leadership of its founder Mani, in whose death 
Kartir doubtless played a significant role; Christianity (see Chapter 18, Part One), which 
was gaining a foothold in Mesopotamia and Iran as early as the second century; and also 
Judaism and Buddhism (see Chapter 18, Part Two). However, Kartir also combated devi- 
ations from Zoroastrianism, which had been open to the influence of Greek philosophy 
and Babylonian astrology during the Hellenistic period. There is therefore no doubt that 
he contributed to the rise of a more orthodox Zoroastrianism, to establishing its supremacy 
over the other religions and, it is thought, to turning it into a 'state religion' linked to the 
power of the throne, though it is not certain that he succeeded in the last objective. 

The Persians had no temples, but worshipped at fire altars on which the symbol of the 
supreme god burned. In the third century, only two types of fire appear to have existed: 
the ddurdn, a lower category established in small localities; and the Vahrdm fires, which 

FIG. 1. Sar Mashhad. Inscription. General view. (Photo: R. Ghirshman. Courtesy of Ph. Gignoux.) 

1 Grenet, 1990; Gignoux, 1991. 



ISBN 978-92-3-103211-0 



FIG. 2. Sar Mashhad. Inscription. Detail. (Photo: R. Ghirshman. Courtesy of Ph. Gignoux.) 

served a province (shahr) 1 and were probably named after the king, as is suggested by 
coins and inscriptions - especially the great trilingual inscription of Shapur I at Naqsh-i 
Rustam (which mentions the fire of Shapur and fires named after members of the royal 
family) - rather than after the god of victory, Vorothraghna? There was also a temple to 
the goddess Anahita, for which Kartir became responsible. 

The Vahrdm fires remained the most important type during the post-Sasanian period, but 
at some moment that is impossible to identify, there came into being three major fires asso- 
ciated with social groups and bearing names that point more to a founder than to a deity. 
Adur-Farrbay was the fire of the priests; Adur-Gusnasp, the fire of the warriors; and Adur- 
Burzen-Mihr, the fire of the farmers. The second of these is the best known to us through 
German archaeological excavations 4 which discovered the exact site at Takht-i Sulaiman 
in Azerbaijan and a large collection of administrative documents, including the temple's 
official seal. The original location of the temple, however, is said to have been Media. The 
location of the other two major fires is also thought to have shifted westwards over the 
centuries, though archaeological research has failed to confirm what may be conjectured 
from some late texts, i.e. that the first of the fires, founded in Khwarizm, is thought to have 
been transferred to Kariyan, in Fars. 5 The site has not been found, but as its name was 
related to the Xvarsnah (royal glory), it was also called Adur- Khvarreh. The third great 
fire was established in Parthia, on Mount Revand (Bundahisn 9, 21), but its site has not 
been identified. It appears to have been the most venerated fire during the Arsacid dynasty. 

2 Harmatta, 1964, p. 226. 

3 Boyce, 1975-78, pp. 222-3. 

4 Osten, von der, and Naumann, 1961; Vanden Berghe, 1979. 

5 Encyclopaedia Iranica, Vol. 1, p. 474. 



ISBN 978-92-3-103211-0 Contents ZOROASTRIANISM 

According to Boyce, 6 this triad of fires reflected the division of Iran between the Parthians, 
the Persians and the Medes. 

Very little is known about the architecture of the fire temples since the only ones to 
have been properly excavated are the large Takht-i Sulaiman complex, the temple of Kuh-i 
Khwaja and the little temple of Tureng-tepe. 7 The monument at Bishapur, identified as a 
fire temple by Ghirshman, 8 is probably not one since it is an underground construction. 
The most common type is the domed chahar-tdq, of which some 50 are known (many 
discovered by Vanden Berghe), consisting of 4 quadrangular pillars connected by 4 arches 
supporting a cupola. They are small monuments with sides less than 10 m long, a few of 
them surrounded by a narrow corridor and some with subsidiary buildings (called aiwans) 
added. This type of monument has often been equated with the Sasanian fire temple but it 
is impossible to say whether the chahar-tdq was a location for public ceremonies, a temple 
reserved for the priesthood or the place where the fire was kept. In addition, it may date 
back to the Parthian period and is certainly attested at the beginning of the Islamic era. 9 
Some of these monuments are hard to distinguish from signal fires that were located at the 
tops of hills or mountains to guide travellers while also serving as places of worship. The 
existence of open temples, however, is highly improbable. 

Although the worship of fire, which had to burn eternally, appears to have constituted 
the central ritual of Zoroastrianism, as is also attested by the Nebeniiberlieferung (Acts of 
the Syriac martyrs, Graeco-Latin sources), the sacrifice of animals accompanied by various 
libations continued under the Sasanians. We learn this from the great Naqsh-i Rustam 
inscription of Shapur, who built up stocks of sheep, wine and bread, in particular for the 
maintenance of pious endowments established for the souls of dead and living members of 
the royal family, an institution which testifies to the importance of individual eschatology 
among the Mazdeans. The sacrificial victim was probably smothered or strangled, but not 
bled as among the Semites. 10 

The doctrine expounded by Kartir, which did not even refer to Zoroaster, was also 
essentially eschatological and could be summed up in a few simple ideas, such as the 
existence of a paradise for the righteous and a hell for the wicked. The need is stressed 
for justice and for obedience to the kings and the gods. Among the latter, explicit mention 
is made of Ohrmazd. In the account of a vision vouchsafed to Kartir by the gods, several 
deities with an eschatological function appear; though not named, they may be Rashn, Mihr 

6 Boyce, 1975-78, p. 473. 

7 Boucharlat, 1987, pp. 51-71. 

8 See Schippmann, 1971, pp. 142-53. 

9 Boucharlat, 1985, pp. 461-72. 
10 Gignoux, 1988, pp. 11-12. 



ISBN 978-92-3-103211-0 Contents ZOROASTRIANISM 

or Vahman. This kind of journey into the other world, although presented for the purpose 
of moral edification and as a warning for the faithful, has several points in common with 
the journey of Arda Viraz, 11 a descent into hell suggesting a shamanistic experience. 

Kartir appears to distinguish between two different rituals, that of Yasna, defined by the 
word yast, and other rituals called kirdagan. 12 It should be noted that liturgical texts had 
not yet been written down: the Avestan alphabet had not been invented and a canon was not 
constituted until the fifth century at the earliest. Until that time, the only written language 
available was Pahlavi, but no written work from before the Islamic period has come down 
to us. 

The Gathas [Hymns] attributed to Zoroaster, and the so-called 'Recent Avesta' texts, 
were not written down until some ten to fifteen centuries later. This came after the invention 
of an alphabet derived from the cursive script of Book Pahlavi, supplemented by extra signs 
and admirably adapted to the phonetics of a language whose geographical home (Khurasan, 
Margiana, Bactria or even Seistan) is still a matter of discussion. Such a precise system may 
even not have been perfected until the Middle Ages by Zoroastrians who had taken refuge 
in India after the Islamization of Iran and who were familiar with the structures of Sanskrit, 
since the Avestan language can be correctly analysed only through comparison with that 
language. Indeed, it is through the Parsees of India that the Avesta manuscript tradition - 
and that of the Late Pahlavi texts - have come down to us. 

The Avestan alphabet 13 comprises over 50 signs, many of them borrowed from Book 
Pahlavi, which serves to represent various phonemes or historical pronunciations. Old 
Avestan is the language of the Gathas (c. 1000 B.C.; divided into 5 sections in verse) and 
of the Yasna Haptanhaiti (in prose), but this arrangement remains rather artificial. The 
Sasanian collection of the Avesta (most of which is lost) and its commentary (called the 
Zand) is described in Book VIII of the Denkard (see page 411). It was composed of 21 
nasks (chapters), divided into 3 sections called gathics (commentaries on the Gathas, and 
the legend of Zoroaster), ritual (liturgy, cosmogony, etc.) and law (primarily juridical texts). 
The extant Avesta contains, besides Yasna 28-53 (= Gathas), various prayers, invocations 
and professions of faith. The yasts (hymns to deities) constitute, together with the Small 
Avesta and the Vendidad (laws relating to purification; comparable to the Book of Leviticus 
in the Bible), the main texts of the 'Recent Avesta'. 

In establishing itself, orthodox Zoroastrianism not only had to vie with Manichaeism, 
which rapidly constituted a body of doctrines presented in a great variety of writings and 

11 Gignoux, 1984a. 

12 Back, 1978. 

13 Hoffmann, 1971, pp. 64-73. 

14 Encyclopaedia Iranica, Vol. 3, Fasc. 1, pp. 36-44. 



ISBN 978-92-3-103211-0 Contents ZOROASTRIANISM 

languages and was thus able to claim universality; it also had to struggle against internal 
deviations, the importance of which has occasionally been exaggerated. Among these devi- 
ations, Zurvanism was probably more a matter of popular belief than of heresy in the strict 
sense. 15 It expressed a more radical dualism than the Mazdean doctrine by making Zur- 
van, the god of time, the father of the two twin gods, Ohrmazd and Ahriman, in this way 
conferring equality on Good and Evil, both of which were thus created by God. This doc- 
trine, which is described in particular in the Nebenuberlieferung, has been given too much 
weight by some authors, who either consider it the normal form of Sasanian Mazdaism 16 
or attempt to expose this heresy in the Pahlavi literature. Zaehner, 17 for instance, believed 
he had detected three strands of Zurvanism, a claim that was rightly rejected by Mole. 18 

Mazdakism, which was as much a social as a religious movement, enjoyed a certain 
success because it was supported by the monarchy under Kavad I (488-531). The move- 
ment was not initiated by Mazdak but by one Zardusht. Thanks to the famous polygraph 
al-Shahrastani, who included a brief analysis of it in his Kitdb al-milal wa 'l-nihal} 9 and 
to a few allusions in the Denkard, its doctrine is quite well known: in some ways reminis- 
cent of communism, the movement impressed people by its teaching that possessions and 
women should be held in common. Shaki 20 has noted its limits, showing how such a system 
of marriage in common or in a group could only have evolved in a very closed society that 
practised incest or consanguineous marriage (xwedodah). In his view, too, the cosmogony 
of Mazdak hardly differed from that of Mazdaism. 21 Recently, however, Sundermann 22 has 
emphasized the Mazdakite community's sense of solidarity in times of famine, and pointed 
out that the consequences of holding women in common would tend to create a matrilinear 
society, which could only offend the Zoroastrians, attached as they were to male predomi- 
nance. The fact that a few women (Boran, Azarmigdukht) attained supreme royal power at 
the beginning of the seventh century may reflect a certain change in attitudes. 

The esoteric or mystical character of Mazdakism, which was welcomed by Arabo- 
Persian authors, who related it to the batiniyya, enables us to define it as a gnosis, the prin- 
cipal traits of which have been defined by al-Shahrastani. Salvation was achieved through 
faith and justice, not through observance. There was no longer any religious obligation for 
those in whom the 4, 7 and 12 powers (an obvious reference to the planets and the signs 

15 See Shaked, 1969. 

16 Christensen, 1944, p. 150. 

17 Zaehner, 1955. 

18 Mole, 1963. 

19 Shahrastani, 1986. 

20 Shaki, 1978. 

21 Papers in Honor of Professor Mary Boyce, 1985. 

22 Sundermann, 1988. 



ISBN 978-92-3-103211-0 Contents ZOROASTRIANISM 

of the zodiac) were united. Shaki recognizes that it was, like any gnosis, a syncretic move- 
ment in which doctrines deriving from the Graeco-Jewish writer Philo, neo-Platonism and 
neo-Pythagorism may also be detected, but that it had little in common with Manichaeism. 
Mazdak is said to have been put to death by Khusrau I (531- 579), who brought the move- 
ment to an end. 

Al-Shahrastani believes that there were a large number of sects or heresies in Sasanian 
Iran. Indeed, as Shaked 23 has shown, a pluralist attitude to faith may have predominated. 
He points to great tolerance over differences in the formulation of doctrines, though such 
deviations from religious orthodoxy did not necessarily entail the constitution of separate 
ecclesiastical structures - it was simply that Sasanian Zoroastrianism was not monolithic. 

The presence and expansion of Zoroastrianism in the eastern provinces of Sasanian and 
Kushano-Sasanian Iran, and further east through Central Asia and as far as China, remain 
obscure, despite large numbers of excavations in regions governed by the former Soviet 
Union. These have so far yielded little evidence apart from archaeological monuments and 
objects relating to the religious or material culture. There are almost no texts apart from 
those found at the famous sites of Turf an and Dunhuang. These do not concern Zoroas- 
trianism but are essentially related to Manichaeism, Nestorianism and Buddhism, which 
cohabited or succeeded one another in East Turkestan. They are fundamental for the his- 
tory of those religions and even more so for our knowledge of Middle Iranian (Sogdian, 
Middle Parthian and Middle Persian) and Turkic. 

Using the literature published in Russian over the last few decades and as a result of his 
own studies, Grenet has shown that the northern regions, Sughd and Khwarizm, provide 
evidence of funeral practices and eschatological doctrines that accord with what we learn 
from the works in Pahlavi (see below) on sixth- and seventh-century ossuaries. For sev- 
eral of these, Grenet 24 has identified scenes of Mazdean liturgy, the representation of the 
six Amesha Spentas, (Abstract Entities), or even the judgment of the soul, in accordance 
with the doctrine of the Pahlavi texts, even though lamentation rites (as on the Panjikent 
frescoes) were normally prohibited. 

In Bactria, Kushan coins attest to a whole series of Zoroastrian deities but to only two 
of the Amesha Spentas, Shahrewar and Wahman, whose relationship to Greek and Indian 
gods provides evidence of a clear tendency to syncretism. 25 The cult of water, concentrated 
on the deified Oxus (Amu Darya), which is represented on these coins as a bearded man 
with a fish, is also found at Panjikent. On the Kushano-Sasanian seals, it is Mithra who 

23 Shaked, 1987. 

24 Grenet, 1986. 

25 Grenet, 1988. 



ISBN 978-92-3-103211-0 Contents ZOROASTRIANISM 

invests the king (further west this falls mainly to Ohrmazd) and plays an important role 
together with the river god. 

Central Asia seems to have been unaffected by the development of the fire cult as 
the only legal form of Zoroastrianism, the type found in western Iran. This is shown by 
the Kushano-Sasanian coins and by the name of the temple in Sughd (va yn < *bagina: 
'dwelling-place of god'), where there appear to be signs of the worship of statues of Iranian 
deities. Of the temples excavated, only Temple B of Surkh Kotal may be regarded as a gen- 
uine fire temple. Those at Toprak-kala (Khwarizm, fourth-fifth century), Er-kurgan (Sogdi- 
ana, second-seventh century, and where a dakhma for exposing the dead has, it seems, also 
been identified), Kurgan-tepe (Samarkand, third-fourth century), Panjikent (fifth-eighth 
century) and Kayragach (Ferghana, fifth-sixth century), among others, are not fire tem- 
ples, although their functions as places of worship are well established. The temple of the 
god Oxus at Takht-i Sangin merely has two 'chapels' reserved for the worship of fire. The 
cassolettes for incense or for the burning of offerings, found in connection with a deity 
in Kushan Bactria, in Sogdiana and as far as Gandhara, should not be confused with the 
receptacles for fire in the fire temples. The lack of texts rules out a detailed overview but 
the essential fact that emerges is that the cult of images was not, in eastern Iran, subject to 
the iconoclastic taboo of Sasanian Zoroastrianism. 

In Margiana and Bactria, Sasanian Zoroastrianism had to compete with Buddhism, and 
also with Hinduism and the worship of Shiva, but the Buddhist complex at Kara-tepe was 
abandoned in the fourth century (?), probably because of the Zoroastrians. Merv had a 
Nestorian community in the fourth century and there was a bishopric at Samarkand, where 
Jews and Manichaeans also lived, in the sixth century. To the north in Khwarizm and Sog- 
diana, on the other hand, Buddhism made little headway. 

As far as is known, relations between Iran and China did not really start until the end of 
the fourth century when, with the decline of the Kushan Empire, the Sasanians were able to 
gain control over the first portion of the Silk Route, the rest of it being in the hands of Sog- 
dian and Indian merchants and later, after the fall of the Hephthalite Empire, of Persians. 
These last did not really establish themselves in China until some time during the sixth cen- 
tury, as attested by the discovery of several hoards of Sasanian coins. There was at that time 
a considerable penetration of Iranian culture into China from Central Asia. One eloquent, 
albeit rather late (784) piece of evidence is the Pahlavi-Chinese tomb inscription of Xian. 26 

Although we are unable to determine the nature of Zoroastrianism in China itself, the 
works in Pahlavi, though not written until the ninth and tenth centuries, are a rich source of 
information concerning the religion as it was practised in the east. This collection includes 

26 Humbach and Wang, 1988. 



ISBN 978-92-3-103211-0 Contents ZOROASTRIANISM 

translations of Avestan texts (the Yasna, the VendTdad, etc.) and compendiums covering 
a wide range of subjects. For example, the cosmogonic myths about the creation of the 
world, together with eschatological doctrines concerning the destiny of the individual soul 
and the events of the last days leading to the restoration of a world identical to that of the 
beginning of time, are found chiefly in the Bundahisn [The Primal Creation] and in the 
VizTdagTha-T Zadsparam. The latter is an 'anthology' which includes a large section on the 
hagiography of Zoroaster, describing the attacks starting in childhood and from which he 
escaped, and his miraculous powers. The same stories appear in the Denkard (Books V and 
VII). Another topic concerns the rules of legal purity, dealt with at length in the VendTdad or 
Sayist-ne-sayist (what is permitted and what is prohibited, with a list of the fines to be paid 
as reparation for transgressions) or the Dadestan-I denig. The Madayan-T hazar dddestdn, 
a document of a more strictly legal nature, provides specific information on family law, the 
law of property and judicial procedures. Other texts concerned with correct ritual are the 
Nerangestan, which deals with liturgical expressions and headings, and the Herbadestan, 
which is chiefly a manual for the use of Mazdean priests. 

Manushcihr, the brother of the ninth-century priest Zadsparam, left three long letters 
in which he protests against the tendency to simplify the ritual, necessitated by the diffi- 
cult conditions in which the community found itself after the Islamization of Iran. These 
conditions are described in numerous texts - though the Muslims are almost never men- 
tioned - in particular, in the Denkard, an enormous compilation divided into nine books 
(the first two of which are lost) and intended to encompass all the knowledge of the Zoroas- 
trians. This work makes much greater use of philosophical argumentation than do the other 
books, frequently using concepts borrowed from Greek or Indian science, in order to refute 
the tenets of Islamic philosophy, as de Menasce 27 has clearly shown, and hence belongs 
to the literature of apology. Faced with Muslim attacks concerning the problems of divine 
attributes and of divine causality with regard to evil, the Mazdeans wanted to defend their 
dualism and reject the patent determinism of certain schools of Islamic thought. The Skand 
gumanig wizar 2 * is even more of an apologia for Mazdaism, not only against Islam but 
also against Manichaeans, Jews and Christians. 

The andarz (wisdom literature), later highly valued by the Persians, is represented in 
the Denkard (Book VI) 29 and in a few other scattered texts. The Mazdean Apocalypse was 
probably a late work, 30 and the events it predicts may be related to the troubles arising 
from the rebellion of Bahram Chobin (sixth century) and even more closely to the revolt of 

27 Menasce, de, 1958; 1973. 

28 Menasce, de, 1945. 

29 Shaked, 1979. 

30 Gignoux, 1986. 



ISBN 978-92-3-103211-0 Contents MANICHAEISM 

Mazyar (between 823 and 840), reported in the Zand-T Vahman yast (a so-called translation 
of an Avestan yast which never existed). 

The royal ideology - immortalized by Sasanian sculpture as its sole artistic expression, 
and associated with the glory of the ancient mythical Kayanian kings - occurs in numerous 
texts. It must have been the basis for the Khwaddyndmag [The Book of Lords], now unfor- 
tunately lost, which linked the history of legendary kings with that of historical kings and 
inspired the authors of epics, the most famous of whom is Firdausi. The alliance between 
royalty and religion, presented as two sisters or two twins unable to live without each other, 
is a recurring theme which, however, must not be taken literally. It seems to have repre- 
sented an ideal which impressed Arabo-Persian writers and provided the inspiration for a 
literary theme. 31 

The palace revolutions which led to a whole succession of kings after the reign of Khus- 
rau II (591-628), and the conquest of the empire by the Arab armies, must have rapidly 
disrupted the Mazdean Church. Wealthy supporters of the temples probably preferred to 
support the new power and Islam in order not to be ruined by the heavy taxes henceforth 
imposed on minorities, since the Mazdeans were held in even lower esteem than the Jews 
or Christians. The priests, as can be seen from certain chapters of the Dddestdn-T denig, 32 
faced serious financial difficulties in celebrating their offices. Many Zoroastrians later emi- 
grated to India, where they are known as Parsees. Here they established flourishing com- 
munities and were able to safeguard the texts of the Sacred Canon which can still be read 

Part Two 

(8. A. Litvinsky) 

The founder of the Manichaean religion, Mani (216-274 or 277), was able to propagate 
his teaching without let or hindrance in Iran during the reign of Shapur I. After Shapur's 
death in 272, however, the opposition of the Zoroastrian priesthood became increasingly 

31 Gignoux, 1984ft. 

32 Kreyenbroek, 1987. 



ISBN 978-92-3-103211-0 Contents MANICHAEISM 

active. Finally, the prophet of the new religion was imprisoned and tortured to death and 
Manichaeism became a persecuted religion in its birthplace, the Sasanian Empire. 

Mani, who was acquainted with many religions, especially Zoroastrianism, Christianity, 
Mandaeism and Buddhism, proclaimed the foundation of a world religion to replace all 
existing religions. He borrowed much from their doctrines and practices, and therefore, for 
all its idiosyncrasies, Manichaeism is a syncretic religion. In consequence, it was partly 
familiar to Zoroastrians, Christians and Buddhists and the prophets of these religions were 
adopted by Manichaeism as its precursors. This encouraged the spread of Manichaeism 
and made it a universal religion, one that was easily propagated among the followers of 
other faiths. 

The basic principle of the Manichaean religion was an all-embracing dualism, reflect- 
ing the dualism of the environment in an idealistic form. In working out his dualistic con- 
ception, Mani borrowed much from Iranian religion and the Gnostics. According to the 
Manichaean religion, the ' two principles' (do-bun), Good (= Light) and Evil (= Dark- 
ness), have existed from the beginning; they are uncreated and are the direct opposites 
of each other. The modern world as a whole, and man in particular, is a mixture of the 
two principles. At the same time, they are not equal - Good is the higher. In the world of 
light, the Father of Light (in Middle Persian, pydr rwsri) (also called the First Parent) sits 
enthroned. The Divinity of Light has a twelvefold Diadem of Light and is surrounded by 
twelve sons. 

The good god is called the God of Light and acts more as an abstract principle than 
a personality. The God of Light has five 'light elements': Ether, Air, Light, Water and 
Fire. His kingdom reaches out limitlessly to the north, west and east. 33 In the south it 
borders on the kingdom of the Prince of Darkness (in Middle Persian, Ahriman). This 
kingdom is divided into five worlds (or caves). The Kingdom of Darkness is the source of 
eternal disturbance and agitation. Incursions are made from it into the Kingdom of Light. 
In outward appearance, the Prince of Darkness is a mixture of a number of different beings. 
34 The origin of the universe is the junction or merging of the forces of light and darkness. 

When the forces of the Prince of Darkness invaded the Kingdom of Light, the King of 
Light offered resistance. For this purpose he evoked two spiritual principles, or rather he 
'called up' the Mother of Life, who in her turn called up Primordial (or the First) Man, her 
son (in the Central Asian texts, Ohrmazd, or Ahura Mazda). He entered the struggle and 
suffered a fatal defeat. The second part of the cosmogonic act was the liberation of captive 
Primordial Man. Although he was liberated, he left five elements, his sons, in captivity 

33 Fihrist, 1970, pp. 786-7. 

34 Ibid., p. 778. 



ISBN 978-92-3-103211-0 Contents MANICHAEISM 

when he returned to the heavens. Nevertheless particles of darkness penetrated into the 
light-bearing principle. With the aim of 'purifying' it, the Father of Light created the visible 
world: the earth, heaven and the heavenly bodies. 

The earth and heaven were created from the bodies of the slain demons, 'from the 
generation [or line] of darkness' (Augustin, Contra Faustin, XX, 9), and the heavenly 
bodies from the liberated particles of light. In order that victory might be complete, a third 
messenger came into the world at the behest of the Father of Light - the Living Spirit. 
He had the attributes of the Iranian sun god Mithra (the name by which he is called in the 
Middle Persian work the Shdbuhragdn) and at the same time he prefigured Mani himself. 
The third messenger evoked the Maiden of Light (Twelve Maidens) and forced the demons 
of darkness to ejaculate the seed and embryos from which plants and animals were born. 
In answer to this, the forces of darkness gave birth to Adam and Eve in the form of gods, 
in an attempt to keep back particles of light so that the flesh would continue to absorb light 
elements. They passed on to the first human couple the particles of light which had been 
absorbed by the forces of evil that gave birth to them, and formed the souls of this couple. 
This was a dark or material soul, made up of such things as lust, greed and envy. Jesus, 
evoked by the Father of Light, aroused this couple, and they became 'new' people instead 
of 'old', able to distinguish between good and evil. 35 At this time a constantly operating 
mechanism was set up for the liberation of the light particles, which finally rose up into the 
Kingdom of Light. 36 

The elements of light were separated from the elements of darkness with the coming 
of Mani. Their complete separation will occur when the Last Judgment takes place - the 
Great War, when the spirit is freed from the body, the particles of light rise up to heaven 
and the carriers of darkness are cast down. 37 It will be a sign of the end of the world when 
the forces of evil prevail and a considerable part of the light is driven out of the world. This 
will be followed by the second coming of Jesus, who will sit in judgment and separate the 
righteous from the sinners. Then heaven and earth will collapse, a Great Flame will arise 
and all the particles of light will be liberated. A new paradise will come into being and evil 
will be fettered and incarcerated in a great stone: the Kingdom of Light will have arrived. 38 

Thus man is destined to work for the liberation of the light particles in his own being 
and in the world around him, and for their union with the principle of light; he must support 
the principle of light. Man should not kill his fellow men, nor should he kill animals; he 
should lead a moral life. Nevertheless, evil and the forces of darkness are very active, and 

35 Boyce, 1975, p. 7. 

36 Fihrist, 1970, p. 782. 

37 Polotsky, 1935; Puech, 1949; Widengren, 1961; 1983; Klima, 1962. 

38 Boyce, 1975, p. 8; for a detailed description in the Shdbuhragdn, see Mackenzie, 1979, p. 513. 



ISBN 978-92-3-103211-0 Contents MANICHAEISM 

man, because of his origin, is predisposed to evil. Man, who has free will, should struggle 
resolutely with evil (darkness) and choose good. 

These precepts were addressed to the Manichaean laity, the ' hearers', both men and 
women, who had to observe certain rules of conduct. For example, they were forbidden to 
kill animals, and they could eat only the flesh of animals that had died a natural death or 
been killed by others. 39 The main aim of their life was to do good and above all to provide 
food for the ' elect' and to serve them. The task of the 'elect' was to pray and to spread the 
doctrine. In addition, they ate vegetable food (especially melons and cucumbers, in which 
light was supposedly concentrated) and bread, and let these foods pass through them. Thus 
their bodies served as a 'filter' for the liberation of the particles of light contained in the 
plants, which then 'along the Column of Glory mount from Earth to Heaven'. 40 Although 
the 'elect' were allowed to eat only vegetable food, they were forbidden to pick the plants 
and fruits themselves - this was the task of the 'hearers'. The preferred drink was fruit 
juice. All sexual life was forbidden. 

The 'elect' who lived in the monasteries were required to journey on foot, spreading 
the doctrine. They ate once a day, after sunset, and were allowed to possess no more food 
than was necessary to feed them for one day and no more clothing than was needed for one 

A pious layman 'hearer' could become an 'elect' and then attain paradise, but only after 
a cycle of reincarnations. (Mani's doctrine included a belief in the reincarnation of souls 
similar to the Indian concept of samsara.) The practice of repentance and absolution was 
widespread. An adherent could acquire the chance of salvation by devoting his life to the 
service of the 'elect', in which case he could hope to be born again as one of the 'elect' 
and attain paradise. The impious would go to hell. The lowest rank in the Manichaean 
hierarchy was that of the 'hearers', then came the 'elect'; even higher were 360 Elders, 
then 72 Bishops and finally 12 Teachers. At the head stood Mani's successor, whose seat 
was in Babylon. 41 

Manichaeism contained a considerable element of social criticism, declaring the world 
to be the incarnation of evil. Manichaeans rejected everything in the world around them, 
including social institutions, and this could be regarded as a form of social protest. Nev- 
ertheless, since they held that evil was eternal, they did not believe it was possible to 
destroy evil on earth. This explains their characteristic pessimism, their feeling that there 
was no solution. According to this doctrine, the rich would inevitably go to the Kingdom of 

39 Asmussen, 1975a, p. 27. 

40 Gershevitch, 1980, p. 282. 

41 Asmussen, 1975a, pp. 29-31; for a detailed analysis, see Schaeder, 1934, pp. 11-16; Van Tangerloo, 



ISBN 978-92-3-103211-0 Contents MANICHAEISM 

Darkness. The future of the adherents of the religion was seen as a liberation from all adver- 
sities - the Kingdom of Light. The element of social protest inherent in the Manichaean 
religion made it attractive to the oppressed masses. 42 

Manichaeism inspired a rich literature. Mani and his followers used a special 'Manichaean 
script' (supposedly invented by Mani himself), which was related to Syriac Estrangelo and 
even more closely to the Mandaean script. Mani used the East Aramaic language, for rea- 
sons which are clear: it was the language spoken over the widest area and could be used as 
a linguistic medium for the propagation of the new doctrine. All the earliest Manichaean 
religious works, with the sole exception of the Shabuhragan, were written in East Aramaic. 
Mani himself is said to have thought and spoken in Aramaic. 43 

For the study of the Manichaean doctrine, the Manichaean sources themselves are of the 
first importance, especially the seven canonical texts: the Shabuhragan, the Living Gospel, 
the Treasure of Life, the Pragmateia, the Book of Mysteries, the Book of Giants and the 
Letters. Only the first of these, the earliest work, is written in Middle Persian; the others 
were compiled in Syriac. Among the other Manichaean sources, the Kephalaia is impor- 
tant. Mention should also be made of the Homilies and a collection of illustrations of the 
most important aspects of the doctrines, the Ardhang. 44 

In addition, in Manichaean circles in Egypt and East Turkestan, canonical Manichaean 
works were translated and original works written in various languages. Many Manichaean 
works have been found in East Turkestan; they are written in Middle Persian, Parthian, 
Bactrian, Sogdian, Ancient Turkic and Chinese. For example, there was a translation of 
the canon into Parthian by Mar Ammo (see below), an associate of Mani. There was an 
extensive Manichaean religious literature in Parthian, consisting of two cycles of hymns, 
prose works about the life and activities of Mani and his death, expositions of the doctrine 
and liturgical texts. 45 The canon was also translated into Sogdian, as were prayers, pre- 
cepts and Manichaean prose works, including the history of the spread of Manichaeism 
in the east, and the text of the confession for the use of the leaders of a Manichaean 
community. Even letters have survived, addressed to various people, including a certain 
Manichaean teacher. 46 Old Turkic works include the penitential prayer of the Manichaeans, 
the X u astvanift. 

Much information (some of it valuable) can be found in Christian polemic works such 
as the Acts of Archelaus {Acta Archelai), the works of St Augustin, the Syrian Christian 

42 Kats, 1955; Sidorov, 1980. 

43 Schaeder, 1934, p. 11. 

44 Boyce, 1968, p. 70. 

45 Boyce, 1954; 1968; 1975; Asmussen, 1975&; Sundermann, 1973; 1979; 1981. 

46 Asmussen, \915b. 



ISBN 978-92-3-103211-0 Contents MANICHAEISM 

chronicles, and the Book of Commentaries of the Syrian Theodor bar Konai. Especially 
valuable Islamic sources are the Chronology of al-Biruni and al-Fihrist al- c Ulum by Ibn 
al-Nadim. 47 

Mani's teaching spread far and wide even during his lifetime and he himself undertook 
a missionary journey to Sind. In the course of his journey, he also visited the town (and 
region) of Turan in Baluchistan. Here Mani was received as a true Buddha, and as a result 
of his preaching many were converted, including the ruler Turanshah and the nobility. 48 
Although Mani claimed that he 'converted the whole country of India to the doctrine', in 
fact he meant the north-west part of India. 49 It is possible that he made other journeys to 
the east - the sources mention 'the Parthian state', 'Kushan' and other lands, 50 but this may 
be a reference to missionary travels undertaken by his disciples. Manichaean missionaries 
engaged in polemics with other religions. According to one Middle Persian text, they were 
especially vehement in their opposition to 'idols, idol priests, altars and their gods'. 51 

Mar Ammo was one of the main preachers of Mani's teaching in the east. He had a 
good knowledge of the Parthian language and script. On Mani's instructions, and in the 
company of the Parthian prince Ardavan and several scribes, Mar Ammo travelled to Abar- 
shahr (Nishapur), whence he continued to Merv. According to the Missionary History, 'he 
ordained numerous kings and rulers, grandees and noblemen, queens and ladies, princes 
and princesses. . . He completed and fulfilled all orders and injunctions that [had been 
given] him by [Mani].' He then travelled further east into the lands of the Kushan Empire, 
where he set up a Manichaean community in one of the towns. 52 

Merv became one of the main centres of Manichaean propaganda in the east. After the 
death of Mani, the head of the Manichaean hierarchy came to Merv, where he found that 
'all the brothers and sisters lived in piety'. He sent one of them, called Zurvandad, with two 
sacred books to Mar Ammo, who was preaching in the town of Zamb (according to Arab 
geographers, this was Zamm on the Amu Darya). In his accompanying letter, he wrote that 
other copies of these books would be made in Merv. 53 

In the east, in Middle Asia and later in East Turkestan, a distinct Manichaean sect 
arose. According to Ibn al-Nadim, it was called 'Dinawariyya', from the Middle Persian 

47 Fliigel, 1862; Hegemonius, 1906; Alfaric, 1918-19; Polotsky, 1934; 1935; Save-Soderbergh, 1948; 
Widengren, 1961, pp. 77-86; Klima, 1962, pp. 401-512; Abel, 1963; Rudolph, 1974; Adam, 1969; Heinrich 
and Koenen, 1975; Asmussen, 1975a, 1975ft. 

48 Henning, 1977, 1, p. 385; Sundermann, 1971ft, pp. 375-6. 

49 Sundermann, 1971a, pp. 88-91. 

50 Henning, 1977, l,p. 386. 

51 Asmussen, 1975ft, p. 13. 

52 Henning, 1977, 1, pp. 200-3; 2, pp. 225-30. 

53 Ibid., l,p. 285. 



ISBN 978-92-3-103211-0 Contents MANICHAEISM 

dendvar, meaning 'giver [or carrier] of religion'. This sect - which constituted an inde- 
pendent church, having its own leader - came into being very early, either during Mani's 
lifetime or soon afterwards. 54 

The Eastern Manichaean Church used Middle Persian as its main sacred language. 
There were also Manichaean texts in Parthian; the Sogdians copied them, and added ver- 
sions in Sogdian. Most of the Manichaean texts discovered in East Turkestan were the work 
of copyists in Sogdian Manichaean communities. Parthian was apparently supplanted as 
early as the fifth century, and was preserved in the Eastern Manichaean Church as a dead 
language. In the sixth century it was replaced by Sogdian in the region of Transoxania, 
although it remained in use in northern Khurasan. There were flourishing Manichaean 
communities in Merv and Balkh. 

Buddhism and Manichaeism coexisted in Central Asia for a long time. Even the most 
ancient Parthian Manichaean texts (poems which can be attributed to Mar Ammo himself) 
contain some Indian Buddhist terms and the number of these increases in fourth-century 
Parthian texts. A Manichaean text on magic, which was probably written on the border of 
Iran and India, perhaps in Balkh, in the sixth century, indicates that there were very close 
contacts between Manichaeans and Buddhists. Sogdian Manichaean texts also include bor- 
rowed Buddhist terms and concepts connected with Buddhist tradition. 55 

Buddhism had a considerable influence on the pantheon, the terminology and even the 
concepts of Eastern Manichaeism, and also on its religious practice. For example, one of 
the central concepts of Eastern Manichaeism, that of the confession of sins, may have been 
borrowed from Buddhism or the reverse. 56 Under the influence of Buddhism, Manichaean 
monasteries (mdnistdns) appeared in the east and later in the west. Information about them 
comes from documents found in East Turkestan, including one Old Turkic document of the 
tenth or eleventh century which mentions a monastery in Turkestan. The document even 
lists the provisions which were to be delivered: 'Every day 30 melons must be taken to a 
big monastery and 30 to a small monastery.' 57 

Manichaean worship included prayers, the singing of hymns, and preparations for the 
feast of the remission of sins. The 'elect' had to pray seven times a day and the 'hearers' 
four times. They prayed facing the sun during the daytime and facing the moon at night. 
Manichaeans had a religious ceremony which marked the imprisonment and death of Mani; 

54 Ibid., 1, p. 202, footnote 1; Sundermann, 1974, pp. 12-128, 131. 

55 Henning, 1977, 1, p. 383; 2, pp. 227-30, 283-4; Asmussen, 1965, pp. 136-47; Sims-Williams, 1983, 
pp. 132-40. For general issues concerning the relationship between Manichaeism and Buddhism, see Ries, 

56 Asmussen, 1965. 

57 Zieme, 1975, pp. 332, 334, 336. 



ISBN 978-92-3-103211-0 Contents MANICHAEISM 

it took place in spring and lasted the whole month. On the last day of this month (i.e. the 
thirtieth) there was the celebration of Bema ('the Throne'). An empty throne was set on 
a dais and a portrait of the prophet was placed on it. 58 The dais had five steps and was 
covered with rich draperies. 

In the second half of the sixth century the Central Asian Manichaean community, then 
led by Shad-Ohrmazd (from Babylon), declared its independence. In the eighth century, 
however, the schism was healed, and during the period of office (710-740) of Mihr as head 
of the Babylonian community his jurisdiction was recognized also in Central Asia. 59 

Manichaeism played an important role in the ideological life of Central Asia right up 
to the time of the Arab conquest. In the year 719, for example, Tokharistan was among 
the countries that sent envoys to China. The embassy was headed by the Great Mu-shia, a 
Manichaean 'elect', and a man profoundly versed in the 'configurations of the heavens'. 
His name is a Chinese transcription of the Manichaean term mocag (literally, 'the teacher'). 
The king of Chaganiyan (a region in Tokharistan) asked the Chinese emperor to confer with 
his envoy on the subject of the condition of the state and 'our religious teachings'. 60 

Following the Arab conquest, a more lenient attitude was adopted towards thebreak 
Manichaeans and many of them returned to Iran and Mesopotamia. Under the c Abbasids, 
however, savage persecutions began again. The head of the Manichaean Church lived 
in Baghdad until the tenth century, when his residence was transferred to Samarkand. 61 
According to Ibn al-Nadim, during this century there were Manichaeans in Samarkand, 
Sogdiana and especially Tunkat (in the Tashkent region). 62 

There were many conversions to Manichaeism in East Turkestan. Its advance was encour- 
aged when the Uighur ruler Bogii kaghan adopted Manichaeism in 762, after which it 
became the established religion of the Uighur state. According to a Uighur text, the ruler 
decided to become a convert after an inner conflict. After two days and nights of unceas- 
ing preaching by Manichaean missionaries, Bogii kaghan appeared before the assembly of 
the Manichaean 'elects', and 'falling before them on his knees and bowing, begged them 
to absolve him of his sins'. Then he said, 'When your "elects" give the command, I shall 
move [act] according to your words and your advice.' 63 Under normal circumstances, how- 
ever, as can be seen from the Chinese version of the Karabalgasun inscription, 64 conver- 
sion required a long time and the participation of the highest echelons of the Manichaean 

58 Boyce, 1975, p. 3. 

59 Ibid. 

60 Chavannes and Pelliot, 1913, pp. 152-3, 197; Belenitskiy, 1954, pp. 44-5; Schafer, 1963, p. 50. 

61 Puech, 1949, pp. 64-5; Boyce, 1975, p. 3. 

62 Fihrist, 1970, p. 803. 

63 Bang and von Gabain, 1972, pp. 32-42. See also Marquart, 1912. 

64 Chavannes and Pelliot, 1913, pp. 190-5. 



ISBN 978-92-3-103211-0 Contents MANICHAEISM 

Church. It is possible that the stream of Manichaeans arriving from Mesopotamia, Iran and 
Central Asia as a result of the Arab conquest and the introduction of Islam contributed 
to Bogii kaghan's conversion. As a result of his adoption of the new religion, the conflict 
between the two schools of Mesopotamian Manichaeism was transferred to East Turkestan, 
where their followers coexisted with members of the Eastern Manichaean Dinawariyya 
community. 65 The Patriarch and the upper hierarchy of the Eastern Manichaean Church 
resided in Kocho, the capital of the Uighur state of Turfan (850-1250). 66 

The importance of Manichaeism in the Uighur state gradually waned. As late as 
983-984, however, the Chinese traveller Wang Yen-te remarked that, in addition to 50 
Buddhist temples there were 'Manichaean temples', and that 'Persian monks pray accord- 
ing to their laws'. Manichaeism had also spread through East Turkestan to China as early 
as 672. Manichaean missionaries appeared at the T'ang court in 694; and in 732 an impe- 
rial edict allowed them to preach 67 to co-religionists, although the same permission was 
not granted to proselytes. Later edicts were more tolerant, but then persecutions began 
again. 68 

Manichaeism played an important role in the development of art. Mani was very fond 
of music, to which his followers ascribed a divine origin (Augustin, De moribus manichae- 
orum, II, V, 16). Even in Mani's lifetime, religious works were adorned with ornaments 
and illustrations to heighten the effect of the text. Splendidly illuminated manuscripts were 
common in the Hellenistic circles with which Mani and his followers were in contact. Mani 
himself was an exceptionally skilled artist (Kephalaia, CLIV, 2). Later, during the Islamic 
era, the name of Mani in Persian literature came to signify an artist of the first order. 
Even Mani's contemporaries were impressed by the size and magnificence of Manichaean 
manuscripts and Arab authors were later to comment on this. The scribes formed a spe- 
cial class among the 'elect'. 69 Texts were usually written in ink on paper, or on silk or 
leather. 70 Manichaean scribes paid special attention to the beauty of their calligraphy and, 
as Boyce states, the manuscripts are written in a 'clear and elegant' script. The headings 
at the beginning of a work and of each section were filled in with vignettes and the text 
was frequently framed with intricate ornamentation. Miniatures, however, were the most 
striking ornament of the manuscripts. They portray all the ranks of Manichaean society - 

65 On the subject of the two schools, see Fihrist, 1970, p. 393. 

66 Sundermann, 1980. 

67 Boyce, 1975, p. 3. 

68 Asmussen, 1975a, p. 24. 

69 Ibid., p. 62. 

70 Boyce, 1975, p. 14; 1968, p. 67. 



ISBN 978-92-3-103211-0 Contents MANICHAEISM 

Mani, the 'elect' (dressed in white) and the laity - and depict religious feasts and symbolic 
images. 71 

The Manichaeans had the reputation of being initiated into magic lore and did, indeed, 
engage in magic. They were, however, also well versed in astronomy (astrology), geogra- 
phy and other sciences. 72 They also produced theologians and poets. 73 

71 Le Coq, 1973; 1979, PL 2-6. 

72 Asmussen, 1975b, pp. 44-5; Boyce, 1968, pp. 75-6. 

73 Fihrist, 1970, p. 803. 







B. A. Litvinsky and M. I. Vorobyova-Desyatovskaya 



Christianity 414 

Indian religions (except Buddhism) 419 

Local religions 421 


Written sources 425 

Inscriptions in Buddhist complexes 432 

Manuscripts from East Turkestan and Central Asia 434 

East Turkestan - a major Buddhist centre 440 

See Map 1. 



ISBN 978-92-3-103211-0 Contents Christianity 

Part One 

(B. A. Litvinsky) 


According to al-Biruni, a Christian preacher appeared in Merv 200 years after the birth of 
Christ, 1 but Christian preachers must have been in Iran even earlier. In the western areas of 
the Parthian Empire, 'Christian communities existed. . . from the beginning of the second 
century, and during the century these communities consolidated themselves by some form 
of organization.' 2 Under the Parthians, religious minorities, including Christians, were tol- 
erated. They gradually spread eastward and the list of bishops of the Syrian Church in 224 
includes the bishop of Dailam, a province to the south of the Caspian Sea. 3 

During the Early Sasanian period the number of Christians increased, but at the same 
time persecutions began. They reached their peak under the mobad Kartir (Kirder) (see 
Chapter 17, Part One), a champion of orthodox Zoroastrianism and of the power of the 
higher Zoroastrian priesthood. In his inscription at the Ka c be of Zoroaster, Kartir records 
the persecution, in c. 280, of Nazarenes (n'cl'y) and Christians ("klstyd'n). During the war 
with the Roman Empire, large numbers of Greek- and Syriac-speaking Christians were 
taken prisoner. These communities, which used different languages, became increasingly 
influential. The head of the Eastern Church was the bishop of Ctesiphon (Seleukia), but the 
other bishops did not always recognize his jurisdiction. 4 At the synod of 410, the bishop 
of the diocese of Abarshahr (Nishapur) was mentioned. 

It can be assumed that the large Christian community in Khurasan came into existence 
as early as 334 in Merv and 430 in Herat. 5 There is some archaeological evidence that 

1 Biruni, 1957, p. 330. 

2 Asmussen, 1983, p. 928. 

3 Sachau, 1915, p. 20. 

4 Asmussen, 1983, pp. 929-32. 

5 Bartold, 1964, pp. 271-2; see also Asmussen, 1983, p. 932. 



ISBN 978-92-3-103211-0 Contents Christianity 

Christian communities existed in Merv and southern Turkmenistan in general from the 
third to the sixth and seventh centuries. The necropolis at Merv, which has several Chris- 
tian tombs, dates back to this time; the ruins of a small Christian monastery and other reli- 
gious buildings have also been found here. A treasure of early Christian gold medallions 
and plaques has been found in Geok-tepe, and impressions of seals bearing a Nestorian 
cross and other objects of a Christian character have been discovered in Ak-tepe (southern 
Turkmenistan). 6 

The history of Christianity in the Sasanian Empire, including Khurasan, was determined 
by three groups of interwoven and constantly interacting factors: the history of the Eastern 
Christian Church itself; the relationship between the Sasanians and the Roman Empire; and 
the political and religious situation within the Sasanian state. A schism took place at the 
oecumenical councils in Ephesus in 431 and in Chalcedon in 451, when the Dyophy sites, 
who recognized a dual nature (the human and the divine) in Christ, separated from the 
main body of the Church. This belief came to be known as Nestorianism, after one of its 
apologists, Nestorius. The Monophysites (who held that Christ had a single, divine nature) 
prevailed and the Nestorians had to flee to the east, where they attained the leading position 
in the Eastern Church. 

Throughout the Early Sasanian period, Christians lived under normal conditions and 
systematic persecutions occurred only when the Zoroastrian priesthood and the state took 
concerted action against them. Even the persecution under Bahram I and Bahram II 
(between 273 and 276), which was linked with the activities of Kartir, 'had no catastrophic 
consequences for the Christian communities, because Narseh (a.d. 293-302) altered the 
directions that had hitherto been followed, giving rise to cool relations with the Zoroas- 
trian dignitaries'. 7 The position changed dramatically under Shapur II (309-379), when 
Christianity became the state religion of the Roman Empire under Constantine. This led to 
the long and systematic persecution (lasting almost 40 years) of their brethren in Sasanian 
Iran. Shapur II's heirs did not continue his religious policy, however, and under Yazdgird 
I (399-421), who was hostile to the Zoroastrian priesthood, circumstances again became 
favourable for the Christians. 

The autonomy of the Iranian Church had been stressed at the third synod, in 424; 8 
the number of Iranian Christians had increased considerably and included members of 
the nobility. Within the Iranian Church a serious struggle then broke out between Nesto- 
rianism and Monophysitism, the latter being more rigid and ascetic than the former. Iranian 

6 Nikitin, 1984, p. 123. 

7 Asmussen, 1983, p. 936. 

8 Ibid., p. 941. 



ISBN 978-92-3-103211-0 Contents Christianity 

FIG. 1. Sasanian Iran. Bas-relief in stucco representing a cross. (Photo: © Musee du Louvre/ Antiquites 

Christians, who had only recently converted from Zoroastrianism, were ideologically closer 
to Nestorianism, which recognized the human nature of Christ and rejected rigid monas- 
tic rules. As a result, the synods of 484 and 497 confirmed Nestorianism as the dominant 
Christian teaching in Iran. This did not mean that Monophysitism was completely ousted, 
however, and under Khusrau II (591-628) it enjoyed the strong support of Shirin, one of 
the shahanshah's two Christian wives, and also of his court physician, Maruta of Tagrit 
(d. 649), who played a considerable role in strengthening Monophysitism. In general, 
Khusrau II showed great favour to the Christians. Shirin enjoyed wide-reaching influ- 
ence and is said to have preached the Gospel in the palace; Khusrau had a church and 
a monastery built for her. There were several Christians among the higher nobility (Fig. 
1). An imperial edict was issued permitting Christians to restore churches that had been 
destroyed. According to al-Tabari, anyone except the magi (members of the priestly caste) 
was allowed to convert to Christianity. 9 A rumour arose, which has survived in Persian 
tradition, that the shahanshah himself secretly became a Christian. 10 

9 Noldeke, Tabari, 1973, p. 287. 
10 Pigulevskaya, 1946, pp. 234^9. 



ISBN 978-92-3-103211-0 Contents Christianity 

Many remarkable people were involved in the work of the Eastern Christian Church. 
One of these was the Catholicos (Patriarch) Mar Aba, a Persian by origin and a former 
Zoroastrian. A man well versed in Zoroastrian lore, he studied extensively and spoke at 
disputations in the Academy of Nisibis (see Chapter 3, Part Two) (where he later taught) 
and in Alexandria, Constantinople and other Christian centres. As a result of the synod 
summoned by Mar Aba in 544, the organization of the Nestorian Church of Iran was more 
clearly defined and became more centralized. 11 The Arab conquest finally brought an end 
to the Sasanian Empire: the last shahanshah, Yazdgird III (632-651), was killed in Merv 
(a Christian outpost in the East and a base for Christian preaching in Central Asia). Inter- 
estingly, it was the Christian bishop of Merv who arranged Yazdgird's burial. 

Although it is known that Christianity came to Bactria ( Tokharistan) from Parthia, 
there is very little information about the Christian community there, 12 and it is unclear 
whether there was a Nestorian bishop in Balkh. 13 The Syrian Book of the Laws of the 
Lands, from the school of Bardaishan (d. 222), gives information about Christian women 
in the 'Kushan country' which, according to Marquart, means Bactria. The statement by 
the fifth-century Armenian author Elishe Vardapet that Christianity had spread to the land 
of 'K'usank c ' and from there southwards to India refers to the reign of Shapur II, that is, 
to the fourth century. In his Christian Topography, Cosmas Indicopleustes writes that the 
Bactrians, 'Huns', Persians and 'other Indians' had many churches. 14 In 549, at the request 
of the Hephthalites, the Catholicos Mar Aba appointed a bishop for all the Christians in 
Hephthalite domains. The History of Mar Aba tells us that later, at the request of the king of 
the Hephthalites and of those Hephthalites who were Christians (Krestyane haptaraye), the 
Catholicos appointed one of their priests as bishop of the kingdom of the Hephthalites. 15 

Christianity was widespread among the Turkic peoples of Central Asia. It is known 
that in 644 Elijah, the Metropolitan of Merv, converted a large number of them - the Turk 
kaghan (king) with all his army 'beyond the river Oxus', that is, in Tokharistan. 16 In 719 
several embassies from Tokharistan went to China. Information drawn from one of them 
shows that Nestorianism existed in Tokharistan, and it is also suggested that there was a 
link between the Nestorian Church and the ruling circles there. 17 An inscription in Si-an-fu 

11 Pigulevskaya, 1979, pp. 204-6. 

12 Sachau, 1919, p. 67. 

13 Spuler, 1961, p. 140. 

14 Bartold, 1964, p. 278, note 88; Mingana, 1925, p. 302; Marquart, 1961, p. 283. 

15 Mingana, 1925, pp. 304-5; Altheim, 1961, pp. 104-5. 

16 Mingana, 1925, pp. 305-6. 

17 Enoki, 1964, p. 72, note 114. 



ISBN 978-92-3-103211-0 Contents Christianity 

(781) mentions a priest named Miles from Balkh. 18 It is known that there were differences 
in dogma and liturgical practice in the eastern regions, including Central Asia. 19 

Christian preachers went from Bactria to Sogdiana, and Syrian sources from 410-415 
provide information about the founding of a Metropolitan See in Samarkand. Even if this 
see was in fact established later, it is still indirect evidence that there were numbers of 
Christians in this region at an early date. Narshakhi reports the existence of a Christian 
church in pre-Arab Bukhara, 20 and according to Ibn al-Nadim there were 'dualists' (i.e. 
Nestorians) and Christians in Sogdiana in ancient times. Archaeological work has con- 
firmed these reports. As one example, a potsherd bearing a fragment of a psalm in Syriac 
has been found in the course of excavations in Panjikent. This was a school text, written 
as a dictation, and judging from the nature of the errors the writer was a Sogdian. It dates 
from the first half of the eighth century, no later than 740. Nestorian burial-grounds have 
also been found in Panjikent. 21 

The Christian mission went from Sogdiana to Semirechye, where archaeological and 
epigraphic evidence reveals the spread of Christianity. A Christian necropolis dating back 
to the sixth or seventh century and an eighth-century Christian church have been found at 
the site of Ak-Beshim. Christian inscriptions in Syriac and Sogdian have also been found 
in Semirechye; and when the Turk tribes came to Semirechye they found Christians there. 
From the epistle of the Nestorian Catholicos Timothy I, it appears that certain Turkic peo- 
ples, probably the Karluks (among whom Christianity was particularly widespread), were 
converted at the end of the eighth or the beginning of the ninth century. One of the most 
famous missionaries of the Nestorian Church at that time was Shubkhalisho, who preached 
in Central Asia. 22 In 893, when Ismail Samani seized the town of Taraz, there was a large 
church there. 23 

Christianity penetrated even further east and reached the Kyrgyz tribes. A runic inscrip- 
tion in Sudzhi mentions the 'instructor in the faith' of the Kyrgyz chief and uses the Syriac 
title mar, which denotes a Christian clergyman. Other Kyrgyz steles with runic inscriptions 
are marked with crosses. 

Christianity penetrated to East Turkestan even before the formation of the Uighur king- 
dom of Turfan. As East Turkestan had very close links with Central Asia, Christian mis- 
sionaries (apparently including many Sogdians) made their way there in the fourth and fifth 

18 Sachau, 1919, p. 68. 

19 Mingana, 1925, p. 321. For more information about Christianity in Bactria (Tokharistan), see Litvinsky, 
1971, pp. 122-3. 

20 Narshakhi, 1954, p. 53. 

21 Paykova and Marshak, 1976; Pay kova, 1979. 

22 Nikitin, 1984, pp. 126-7. 

23 Narshakhi, 1954, pp. 86-7. 



ISBN 978-92-3-103211-0 Contents Indian religions (except Buddhism) 

centuries. They were to play an even greater role in the fifth and sixth centuries, the period 
during which it appears (according to the earliest Christian texts found in East Turkestan) 
that the first Christian communities appeared there. 24 The translation of Christian works 
into Middle Persian in Iran and into Sogdian in Sogdiana had already begun. Many remains 
of Christian texts have been discovered in the sites of East Turkestan, including the library 
of the ruined Nestorian monastery in the oasis of Turfan. They were written in Syriac, 
Middle Persian, Parthian, Khotanese Saka, Sogdian and Turkic. Sogdian texts predomi- 
nate, however, and Sogdian became the second most important language (after Syriac) in 
the Nestorian Church in Central Asia. 25 

According to the Chin-shih (Chapter 124), the family of Ma Si-ling-tsisa (i.e. Mar Sar- 
gis) settled in Lintau (in what is now the province of Gansu) in 578. Christian mission- 
aries clearly penetrated even further east and in 635 Bishop Aloben (A-lo-pen in the Chi- 
nese sources) reached the T'ang court, bringing with him sacred books and images. The 
emperor treated him favourably and a Christian monastery was built, marking the begin- 
ning of Christianity in China itself. 26 

Indian religions (except Buddhism) 

As an important source of religious teachings, India had a major influence on the peoples 
of neighbouring countries. This was true of both Buddhism and Hinduism. During the 
period of the Graeco-Bactrian kingdom in northern India and Pakistan, and apparently 
in Afghanistan as well, the worhip of Vishnu was widespread. Vishnu was regarded by 
his worshippers as the basis of creation and the source of all that exists. 27 Judging from 
certain Indo-Greek coins bearing the image of a humped bull, the worship of Shiva was also 
widespread 28 - he was a fierce god, combining the characteristics of the Vedic Rudra and 
some non- Aryan fertility god. Shiva rode on the humped bull, Nandi, and was frequently 
accompanied by his wife, the beautiful Parvati. The linga, a phallic pillar, was often used 
as a symbol of Shiva. 

There is much evidence that the worship of Shiva flourished in northwestern India 
immediately before the coming of the Kushans. For example, a group of coins of the Indian 
Saka ruler Maues bears the image of Shiva. In Sirkap (Taxila), in a stratum corresponding 

24 Hansen, 1968, p. 93. 

25 Nikitin, 1984, pp. 128-30, including a detailed bibliography. 

26 Kychanov, 1978, pp. 76-8. 

27 Tarn, 1951, pp. 137, 172, 381, 391, 406; Agravala, 1970; Majumdar, 1970, pp. 30-2. 

28 Tarn, 1951, pp. 135-6, 163, 172-3,213. 



ISBN 978-92-3-103211-0 Contents Indian religions (except Buddhism) 

to approximately the first century a.d., a bronze seal has been discovered bearing the image 
of Shiva with an inscription in Brahmi and Kharosthi: Sivaraksita (protected by Shiva). 

The worship of Shiva was also widespread in the Kushan Empire, as can be seen from 
Kushan coins: those of Vima Kadphises, Kanishka, Huvishka and Vasudeva bear a figure 
of Shiva or of Shiva with his vehicle, the bull Nandi. The inscription Oeso on coins refers to 
Shiva. It used to be believed that this legend reflected the Prakrit word havesa (in Sanskrit, 
bhavesa), an epithet of Shiva meaning 'Lord of Being', or the Prakrit development vesa 
of the Sanskrit vrsa (bull). 29 The theory has now been advanced that this name is in fact 
an East Iranian development of the name of the Zoroastrian wind god Vayu, who took on 
the iconographic appearance of Shiva. 30 A stele has been discovered in Mathura with the 
image of two Kushans worshipping the Shiva linga. 

The worship of Shiva also spread to those areas of the Kushan state that are now in 
Afghanistan, and reached the Amu Darya (Oxus). In Airtam, near Termez, a stone slab has 
been found with a Bactrian inscription and a carved image of Shiva. In Soazma-kala, near 
Balkh, a stone slab was discovered bearing the image of a three-headed standing Shiva with 
a trident and other attributes, which strongly resembles Hercules. 31 A wall-painting still to 
be seen in the temple of Dilberjin has a central group depicting Shiva and Parvati mounted 
on the recumbent bull Nandi, 32 which cannot be earlier than the fifth century. Many works 
of Shivaite art and of Hindu art in general, dating back to between the fifth and the eighth 
century, have been discovered in Afghanistan. These include some 25 marble sculptures 
and other works of art. 33 Recent excavations in Panjikent have yielded a large sculptural 
group of Shiva-Parvati. 

Brahmanism played an important role in the religious life of Afghanistan in early 
medieval times; individual brahmanic images even penetrated into Buddhist circles and 
can be found in Buddhist sites such as Tepe Sardar. Tndianization' also affected Bamiyan. 

All these influences must have affected the religion and (to a greater degree) the fine 
arts of Central Asia. The wall-paintings of Panjikent include images directly related to 
the iconography of Shiva worship, in particular a standing three-headed god. Yet in many 
details they are quite unlike the Indian prototypes. The clothing of the three-headed deity 
bears the inscription wsprkr (or wysprkr). Linguists suggest that this name links the Sog- 
dian Vysprkr with the name Oeso on Kushan coins, which in turn comes from the Zoroas- 
trian wind god Vayu. Ves-parkar (Veshparkar) is believed to come from the Avestan Vaiius 

29 Rosenfield, 1967, pp. 22 et seq., 93-4, 111, etc. 

30 Humbach, 1975, pp. 404-5. 

31 Fischer, 1957. 

32 Kruglikova, 1976, pp. 93—4, figs. 54, 55. 

33 Kuwayama, 1976. 



ISBN 978-92-3-103211-0 Contents Local religions 

Uparo Kairitio ('the wind whose action spreads in the upper region'). In Sogdian trans- 
lations of Buddhist texts, he corresponds to Shiva (Mahadeva) and is described as having 
three faces. Thus this god with a Sogdian name and found in the art of Panjikent appears in 
a form connected with the worship of Shiva, which is in keeping with the Sogdian written 
tradition. 34 

Local religions 

Despite the fact that Nestorianism, Manichaeism and Buddhism spread in the non-nomadic 
areas of Central Asia such as Tokharistan, Sogdiana and Khwarizm, most of the population 
continued to profess the local Iranian religion. This could be described as a Central Asian 
version of Zoroastrianism, which differed substantially from the orthodox Iranian form. 

In Sogdiana, according to Chinese sources, 'they honour the Buddhist religion; they 
sacrifice to the god of heaven', which Chavannes takes as referring to Mazdaism. 35 The 
texts report that believers worshipped a golden image and sacrificed animals to it. Thou- 
sands of worshippers came every day to offer sacrifice. According to other information, 
in Samarkand 'the king and the people did not follow Buddhism, but worshipped fire'. 36 
There was a temple for ancestor worship in the palace of the Sogdian ruler. The feasts and 
customs of the Sogdians are also described (in the eleventh century) by al-Biruni. He men- 
tions that the Sogdians celebrated the coming of the New Year, which was connected with 
ideas about the death and revival of nature. Once a year, the people of Sogdiana mourned 
the dead. When doing this they lacerated their faces and offered food and drink to those 
who had died. 37 

The worship of Siyavush was connected with the worship of the dead. On the first 
day of the New Year a cock was sacrificed to him. It was believed that the divine youth 
had died and his bones had been lost. On a particular day, the believers, dressed in black 
and bare-footed, looked for them in the fields. The custom of burial in ossuaries (ceramic 
receptacles for bones) was widespread in Sogdiana, Khwarizm, the oasis of Tashkent and 
Semirechye. When the flesh had fallen away from the bones, these were gathered into the 
ossuaries, which were placed in a special chamber. Some ossuaries were richly decorated 
with magnificent reliefs. 

Sources relating the Arab conquest mention 'fire temples' and 'idol temples' . They were 
richly decorated and contained many precious objects for use in worship - for 

34 Humbach, 1975, p. 404; Belenitskiy and Marshak, 1976, pp. 78-9. 

35 Chavannes, 1903, p. 135. 

36 Hui-li, 1959, p. 46. 

37 Biruni, 1957, p. 258; see also pp. 236, 255. 



ISBN 978-92-3-103211-0 Contents Local religions 

example, a pearl the size of a hen's egg is mentioned. The imagination of the Arab con- 
querors was fired by the size of the gold and silver idols in these temples. 38 Firdausi's 
Shah-name also mentions 'fire temples' in Bukhara and Paikent. The ninth-century Pahlavi 
geographical treatise Sahrlha-T Eran mentions the establishment of a 'miraculous fire' in 
Samarkand by Siyavush's son and the placing there for safe -keeping of gold (or gilded) 
plates inscribed with the text of the Avesta. The treatise also mentions the destruction, 
by Sokandar (Alexander of Macedon), of these plates, after which the Turanian Frasiak 
(Afrasiab) 'made the dwelling of the gods into temples of the daevas' . In Kushaniya stood 
a temple on whose walls were depicted the ancient kings of various nations. 

Sogdian texts mention various divinities. First there is zrw ( Zurvan). In translations of 
Buddhist texts this name is used in place of the Indian Brahma. He is called 'King Zurvan' 
and given the epithet of 'great' and 'king of gods'. His particular iconographic feature was 
the beard. Xwrmzf fiy( Ahura Mazda, or Orhmazd), the supreme god of Zoroastrianism, 
is very rarely mentioned. There is more frequent mention of A 8j3a y, who held the second 
place in the divine hierarchy and was worshipped by Zrwsc (Zoroaster). Clearly Ohrmazd 
(Ahura Mazda) was also referred to as A 8fia y. The deity Veshparkar (wsprkr) has already 
been mentioned. All this is evidence of the Zurvanite tendency of Sogdian Zoroastrianism 
and of certain links with Indian religions. There is also mention of Vr Ora yna (ws yn), 
Druvaspa ( Srw'sp), Haoma ( ywm), X v armah (prn) and other gods, among whom one of 
the most important was the female divinity Nanai. It is significant that the Amesha Spentas 
(see Chapter 17, Part One) occur in personal names. Henning has described the Sogdian 
religion as 'the impact of Zoroaster's teachings on the native paganism of Sogdiana' 39 - 
perhaps it would be more accurate to say 'the native Iranian paganism'. 

Religious iconography and architecture, as well as burial customs, are proof that the 
local religions of Sogdiana, Khwarizm and Tokharistan were closely related. The princi- 
ples of the ancient Iranian religion, which was dualistic, served as the unifying factor and 
idiosyncratic variations of these principles continued to develop right up to the conversion 
to Islam. At the same time, there were considerable changes in their pantheon and prin- 
ciples, which assimilated many new subjects, elements and figures that did not appear in 
the west Iranian religion. Some of these inclusions were due to interaction with religious 
systems of foreign origin (see also Chapter 17). The religion of the Central Asian nomadic 
Iranian peoples was even more idiosyncratic. From the fourth to the eighth century they 
continued the practice of burying the dead in tumuli. Both inhumation and cremation were 

38 Belenitskiy, 1954; Gafurov, 1972, pp. 285-7. 

39 Henning, 1965, p. 250. 



ISBN 978-92-3-103211-0 Contents Local religions 

The beliefs of the ancient Turkic and Mongol peoples belong to a completely different 
religious and mythological system. Their cosmogonic ideas are known from runic inscrip- 
tions in honour of Kiil-tegin and Bilge kaghan: 

When high above the blue sky was created and down below the brown earth had been created, 
between the two were created the sons of men. When this took place, the heaven rose up like 
a roof above the earth. 

The rising sun was also worshipped and the east was considered the most important direc- 
tion: the doors of the kaghan's tent faced east. The earth was seen as square; the kaghan's 
headquarters was in its centre, where the Turk people lived, while their enemies lived 
around the periphery. There are hints of a myth concerning a cosmic catastrophy which 
was part of a cosmogonic myth. 

The Turkic peoples, and those belonging to the Siberian Central Asian religious system 
in general, saw the universe as divided into three parts: an Upper (= Heaven), a Middle 
(= Earth) and a Lower World. The Lord of the Lower World (the Underworld) was Erklig 
(in Mongolian, Erlik kaghan). One member of his retinue was Burt, the spirit of sudden 
death; he was opposed by Tengri (Heaven), the supreme deity of the pantheon. Tengri 
determined the order that prevailed in the world and also the destinies of people. The West 
Turkic people of the Khazar conceived of Tengri as a hero of gigantic size; tall trees were 
dedicated, and horses sacrificed, to him. According to al-Kashgari, the Turks used the word 
'Tengri' to mean 'high mountains' and 'big trees'. According to Chinese sources, it was 
on high mountains that the Eastern Turk kaghans and 'nation' offered prayers to the 'spirit 
of heaven'. Tengri's divine consort was Umai, the goddess of fertility, protectress of the 
new-born. Prayers and sacrifices were offered to the sacred 'Earth- Water', belonging to the 
Middle World. 40 

In the mythology of the ancient Mongol nomads of the steppe (the Hsienpi, Kitans and 
others), there were two main cosmic principles, Heaven and Earth (cf. the ancient Turk 
mythology). The Hsiung-nu and Wu-huan worshipped the spirits of their ancestors, heaven, 
earth, the sun, the moon and the stars. The legends of the Hsien-pi tribes included zoomor- 
phic elements that had their origins in clan totems: a horse, a bull, a deer (elk). The Kitans 
worshipped abstract gods - which they personified - such as war and fire (sometimes in 
the form of totemic animals - a white horse or a deer) and also the spirits of their ances- 
tors. They believed that the Initial Order, mounted on a white stallion, had met a celestial 
maiden in a cart drawn by a cow near the mountain of Mu-ye. Eight sons, the forefathers 
of the Kitan tribes, were born from the marriage of Initial Order and the celestial virgin. 

40 Roux, 1956-57; 1962; Potapov, 1978; Klyashtorny, 1981. 



ISBN 978-92-3-103211-0 Contents Local religions 

The Mongols themselves had a dualistic system. Their supreme deity was Tengri 
(Heaven), the creator of all that exists, who determined the fates of men and affairs of 
state. He was called blue and eternal, and was viewed as the male principle. His counter- 
part was Utiigen, the goddess of Earth, who was associated with notions of fertility and 
the rebirth of nature. Utiigen was the feminine principle just as Umai was in ancient Turk 
mythology. The Mongols also worshipped the sun, which was regarded as the mother of 
the moon. There were distinctly totemic ideas in their legends. 41 Ancient Altaic religious 
and mythological systems and religious practices were shamanistic. 

The religions of Central Asia not only co-existed, but also interacted, and competed 
for adherents. The religions which penetrated into Central Asia, such as Zoroastrianism, 
Buddhism, Manichaeism and Christianity, had to establish themselves in an environment 
where other beliefs existed, whether Tokharian, Iranian (the Khotanese Sakas, Sogdians, 
Bactrians, Khwarizmians, etc.) or Altaic (ancient Turkic and Mongol). Even the spread of 
the ideas embodied in Zoroaster's teaching did not have the same consequences among the 
eastern Iranians as in western Iran, because a substratum of local beliefs and rituals sur- 
vived. There were considerable differences between Central Asian and Persian Zoroastri- 
anism. In other cases, several religious systems were superimposed, for example in the case 
of the spread of Buddhism, Manichaeism and Christianity among the Turks and Uighurs, 
whose shamanistic beliefs remained, to a greater or lesser degree, the substratum of their 
spiritual concepts. 

As this process continued over many centuries, only a few traces were left of the local 
religion. This was, for example, the case with the Khotanese Sakas, whose conversion to 
Buddhism was so complete that almost nothing remained of their original, ancient Iranian 
religion; the same is true of the Tokharians. The Sogdians, who kept many elements of their 
original religion, were in an intermediate position. There was also interaction between the 
various new religions, both local and imported, which had a developed written tradition. 
This led to borrowings, the transference of religious concepts and ideas, and syncretism. 42 

41 Wittfogel and Feng, 1949; Rintchen, 1959; 1961; 1975; Tucci and Heissig, 1970; Zhukovskaya, 1977; 
Neklyudov, 1981. 

42 Asmussen, 1965. 



ISBN 978-92-3-103211-0 Contents Written sources 

Part Two 


{M. I. Vorobyova-Desyatovskaya) 

Written sources 

As mentioned in previous chapters, the enormous territory encompassing Central Asia and 
East Turkestan constituted a single region in the final centuries B.C. and during the greater 
part of the first millennium a.d. This is explained by its common ethnic character, its 
shared historical fortunes, the similar geographic and economic circumstances that condi- 
tioned the character and pace of its socio-economic development, and the region's cultural 
similarity. A basic factor determining this cultural similarity was Buddhism, which was 
accepted over the entire territory as a doctrine of moral ethics, an ideology and a religion. 
The spread of Buddhism beyond the boundaries of India took place at the time of the 
Mauryan emperor Ashoka (c. 268 B.C.), and found expression in his edicts engraved on pil- 
lars and rocks at various points throughout the empire. In the propagation of Buddhism and 
Indian culture in Central Asia and East Turkestan (of which Ashoka Maury a was the initia- 
tor), an important role was played by the Parthian, Saka and Kushan rulers of north-western 
India. A major influx of Buddhist missionaries into these territories occurred among the 
Kushans. The principal route of Buddhist expansion lay through Bactria and the western 
possessions of the Kushan Empire. Part of this territory (northern Bactria on the right bank 
of the Amu Darya, or Oxus - later Tokharistan) now forms part of Middle Asia (southern 
Uzbekistan and Tajikistan). Recent archaeological findings have shown that Buddhism and 
Indian culture permeated every area of Central Asia, leaving direct evidence in the form of 
inscriptions and religious structures, as well as profound traces in the cultural substrata of 
the local peoples. Buddhism only ceased to play an important role in the region from the 
end of the eighth century, after the arrival of Islam, and in the northern region of Central 
Asia (in Semirechye) Buddhist religious centres were evidently still functioning as late as 
the tenth century. 



ISBN 978-92-3-103211-0 


Written sources 

rw% V mjn&'-^ST** **!^F ! 



FIG. 2. Fragment of a manuscript on paper from the collection of N. F. Petrovskiy (no. SI P/7). 

Text in Sanskrit Saddharma-pundarika. 

On the lower level an inscription in Khotanese Saka. 

Script in South Turkestan Brahml. Coloured illustration: Buddha receiving offerings. 

In the Kushan period, Buddhism also began to be actively propagated in East Turkestan. 
Buddhist penetration followed two paths: from Bactria, the centre of the Kushan posses- 
sions, to Kashgar and further east; and from northwestern India and Kashmir to Khotan 
and the southern oases of East Turkestan (see Chapter 15). Although no precise informa- 
tion is available about the time when Buddhism penetrated to the northern oases of Turfan 
and Kucha, it probably became established there at the beginning of the Christian era. In 
300 Chinese sources report the presence in Kucha of 1,000 Buddhist temples and sanctu- 
aries; and in the fourth century Kucha became an important centre of Buddhist education, 
where translators were trained for China. It was from here that Kumarajiva, the celebrated 
translator of Buddhist texts, whose school was considered one of the most authoritative 
in China, was invited to that country. Judging by the borrowing of some Prakrit Buddhist 
terms from the Khotanese Saka language into the Tokharian languages, Buddhism came to 
the Tokharians from the south, in other words, from Khotan (Fig. 2). 

One of the most important results of India's cultural influence on East Turkestan was 
the dissemination of ancient forms of Indian writing - Brahml and KharosthI - which 
were adopted by the local peoples (who had no script) to set down their languages. From 
the adaptation of Indian Brahml to the writing of the Tokharian A and B languages in 



ISBN 978-92-3-103211-0 Contents Written sources 

the northern oases of East Turkestan (Turfan and Kucha), a version of Brahmi known as 
Central Asiatic slanting Brahmi emerged. 

In Khotan and Kashgar, Indian Brahmi was adapted to the writing of the Saka language 
(one of the East Iranian languages current in the area), which was the tongue of the original 
population. From this adaptation another version of Brahmi emerged - Central Asiatic 
upright Brahmi, the writing of the southern oases. These processes had evidently been 
concluded by the fifth century. Palaeography is today used as one of the criteria for dating 
manuscripts in the Brahmi script and for determining where they were copied. 

Another form of ancient Indian writing which found its way into East Turkestan and 
Central Asia was KharosthI, long thought to be a chancery hand which had spread to 
north-western India from the Achaemenid chancery. The Kushans promoted it to the rank 
of official state script and it was employed in Bactrian territory alongside the local Bac- 
trian writing system. Besides two versions of the Bactrian script, other forms of writ- 
ing which gained currency in Central Asia included Parthian, Sogdian, Khwarizmian and 
Pahlavi (Middle Persian). This is evidently the reason why the Indian scripts, KharosthI and 
Brahmi, did not develop local versions here; and in the following centuries their use was 
confined to a Buddhist context, where they appeared in inscriptions on monastic ceramic 
ware and similar articles, on reliquaries and in Buddhist manuscripts (Fig. 3). 

In East Turkestan, no inscriptions on monastic ware have been discovered, but in the 
first and second centuries, local coins with a 'Sino-KharosthI' legend were being minted in 
Khotan. The largest find of written monuments in KharosthI characters is an archive of offi- 
cial documents on wooden strips, dating from the mid-third to the mid-fourth century and 
found in the southern oases of Niya and Kroraina in the territory of the small, independent, 
ancient state of Kroraina in the oases of the Taklamakan desert. 

The inscriptions on monastic ceramic ware and reliquaries (deposited in Buddhist reli- 
gious buildings) and Buddhist manuscripts are the main written sources that have enabled 
modern research to determine the territorial spread of Buddhism, the time when this prop- 
agation took place, and the principal Buddhist schools and centres which have influenced 
the region's history and culture. The Buddhist manuscripts found in East Turkestan and 
Central Asia remain the only original texts of the Buddhist canon known anywhere in the 
world. Even in India, virtually none of these texts has been preserved. Some of the Bud- 
dhist writings, and commentaries on them, found in India are merely in the form of late 
copies. Questions as to how the doctrinal, philosophical, dogmatic and religious tenets of 
Buddhism came to be established in the territories of India and beyond, and what transfor- 
mation these aspects of Buddhism underwent in the new cultural milieu, can be answered 



ISBN 978-92-3-103211-0 


Written sources 

FIG. 3. Clay pot with an inscription in Brahmi. Found in Afrasiab near Samarkand. 

only on the basis of the information obtained from the manuscripts of East Turkestan and 
Central Asia. 

A large group of Buddhist texts, found in the territory of East Turkestan and in the 
ideologically, religiously, culturally and (for almost a century) administratively associated 
Dunhuang, were written (or copied) in these areas in Chinese, Tibetan and Uighur. For 
copying the Buddhist texts, the Turks, among whom Buddhism started to be professed as 
early as the sixth century, used the Uighur (Sogdian in origin), Brahmi (seventh-eighth cen- 
turies) and Tibetan (eighth-century) scripts. Turkic Buddhist manuscripts and xylographs 
have been found in Turfan, Dunhuang, Miran (south-west of Lake Lop Nor) and Suzhou. 
The territories in which Buddhist texts in Indian scripts have so far been found are listed 
below, together with a brief description of the texts themselves. 


(a) Manuscripts written in Brahmi characters (Indian and slanting), in Sanskrit and in 
Tokharian A and B, have been found in the northern oases - Turfan and Kucha - as 
well as in Dunhuang (for details of the manuscripts, see below). 



ISBN 978-92-3-103211-0 Contents Written sources 

(b) Manuscripts written in Brahmi characters (Indian and upright), in the Khotanese, 
Tumshuq Saka and Sanskrit languages, have been found in Khotan and Kashgar. 

(c) Manuscripts written in Kharosthl characters in Gandhari, in which a local non-Indian 
substratum is traceable, have been found in the oases of Niya and Kroraina to the 


_ kk«nyi ^LjaiFy- Jj~ft#5« Ji^itr: 

tt 1 

■■,■- >«■ 

FIG. 4. Manuscript on birch bark in Kharosthl script with the text of the Dharmapada. Photo: © 
Bibliotheque Nationale de France. 



ISBN 978-92-3-103211-0 Contents Written sources 

south of Lake Lop Nor (around 800 official documents on wooden strips, including 
many texts concerned with the activities of the local Buddhist community and the 
performance of religious rituals). 

(d) A manuscript in the KharosthT script in Gandhan, on birch bark, containing the text 
of the Buddhist Hinayana sutra, the Dharmapada (Fig. 4), has been discovered in 

(e) A small fragment of palm leaf bearing a text in the KharosthT script in Gandhan, and 
evidently an excerpt from the Hinayana version of the Mahaparinirvana sutra, has 
been discovered although the location of the find is unknown. The fragment, which 
has not been published, forms part of the S. F. Oldenburg Collection in the Institute of 
Oriental Studies of the Russian Academy of Sciences in St Petersburg. The fragment 
shows that the KharosthT script was used for making manuscript copies of Buddhist 
canonical texts, and that the Dharmapada find is not an exception. 

(f) The only known dedicatory inscription to a Buddhist community, in Gandhan and 
written in KharosthT characters, was found in Lo-yang. 


(a) In the Kushan period, Old Termez was a prominent Central Asian Buddhist centre 
and many remnants of religious buildings have been found in the region, including the 
Buddhist monastery of Kara-tepe and the Fayaz-tepe monastic site. The Old Termez 
area has provided a large quantity of ceramic inscriptions written in KharosthT and 
BrahmT script connected with the activity of the Buddhist community of 
Mahasamghika. Dedicatory inscriptions on reliquaries and other objects have also 
been found. 

(b) The castle of Zang-tepe, 30 km to the north of Termez, probably had a Buddhist 
stupa. Fragments of various Sanskrit manuscripts written in BrahmT script on birch 
bark have been recovered. 

(c) Round the demolished Buddhist stupa at the palace of Kafyr-kala (in the town of 
Kolkhozabad in the Vakhsh valley), scorched remains of manuscripts in BrahmT script, 
written on birch bark, have been recovered from the debris. 

(d) At Balkh, excavations of the Buddhist complex have brought to light a broken piece 
of pottery with a Buddhist dedicatory inscription written in KharosthT characters. 



ISBN 978-92-3-103211-0 Contents Written sources 

(e) In southern Bactria, a copper vessel has been found with a dedicatory inscription, in 
which mention is made of the Buddhist school of Dharmaguptaka. 


In Merv, archaeologists have discovered an important Buddhist centre, with temples and a 
number of stupas. In one of the stupas in Merv, and in another stupa opened by chance in 
Bairam-Ali, immured vessels with two Buddhist manuscripts have been found: both were 
written in Sanskrit in BrahmT characters and inscribed on birch bark. One of them has been 
restored and is preserved at the Institute of Oriental Studies in St Petersburg; the other is 
undergoing restoration in Moscow. 


In the Krasnorechensk site, in the Chu valley, excavations have been in progress for many 
years on the site of the Buddhist temple. In 1985, in the passageway not far from the 11- 
m-high statue of Buddha in Nirvana, manuscript fragments in Sanskrit, written in BrahmT 
characters on birch bark, have been found. 


As regards contacts between Iranian and Indian cultures, 43 the numerous finds of Buddhist 
relics - monasteries, stupas and rock inscriptions - show that Buddhism flourished in the 
Kushan period. Ceramic, reliquary, vase, wall and rock inscriptions in Kharosthl script have 
been discovered in the following areas: Wardak, Begram and Gul Dara (all near Kabul); Tor 
Dheri (in Pakistani Baluchistan); and Dasht-i Nawur (a trilingual rock inscription in Bac- 
trian, Kharosthl and an undeciphered script similar to Kharosthl). Manuscripts in BrahmT 
and Kharosthl characters have also been found in the Bamiyan monastery (Afghanistan). 


The diffusion of Buddhism in this territory over a lengthy time span, from the Kushan 
period to the Muslim conquest, is attested by the many rock inscriptions left by Buddhist 
pilgrims and by reliquary inscriptions from Bajaur, Tirah, Swat, Gilgit and Hunza written 
in Kharosthl and BrahmT scripts. In the first millennium B.C., the Gilgit and Hunza areas 
were crossed by Buddhist pilgrims, merchants and traders travelling between Central Asia, 

43 Identified by Fussman, 1989, p. 445. Thanks to Fussman's brilliant work, devoted to the palaeography 
of inscriptions in Indian scripts and to the Gandhari language, the study of Indian inscriptions can, for the 
first time, be placed on a scientific basis. 



ISBN 978-92-3-103211-0 Contents Inscriptions in Buddhist complexes 

India and East Turkestan. It was in Gilgit, in the early 1930s, that Sir Aurel Stein made 
one of the most outstanding discoveries of the century - a Buddhist library of the seventh 
century, immured in a stupa and comprising Buddhist manuscripts, on birch bark and paper, 
written in Sanskrit in Brahmi characters. Most of the manuscripts relate to the canon of the 
Mulasarvastivada Hinayana school. 


This area is rich in Buddhist monuments of the Kushan period. Many inscriptions writ- 
ten in KharosthT script have been found on Buddhist reliquaries, vessels and walls in the 
monasteries of Bimaran, Hadda, Jalalabad and Basawal. 

Inscriptions in Buddhist complexes 

The contents of the inscriptions listed above provide information about the geographic 
spread of Buddhism and its various schools; about the chronological sequence of the dif- 
fusion process; and about the names of those who professed the religion. In some cases, 
this information can help to establish the ethnic group or native language of the donor or 
proprietor of the object, or the social status and occupations of the followers of Buddhism. 

Chronologically speaking, Buddhist inscriptions in KharosthT characters encompass the 
period from the first century B.C. to the second century a.d. (the earliest attested dating is 
58 B.C. and the latest a.d. 129). Inscriptions in Brahmi characters, on the other hand, bear 
no dates. Through palaeography and an analysis of the accompanying material, they may 
be dated approximately to the fourth-fifth centuries a.d. 

It is legitimate to regard the oldest groups of inscriptional finds as the most represen- 
tative. To judge from the finds made in Buddhist monasteries and temples, dedicatory 
inscriptions were usually made on earthenware that was donated to the Buddhist com- 
munity. Another type of earthenware inscription is found on articles belonging to people 
living in the monasteries. In the area of Middle Asia (Uzbekistan, Tajikistan, Turkmenistan, 
Kyrgyzstan, etc.) where a number of monastic complexes have been excavated - Kara- 
tepe, Fayaz-tepe, Ajina-tepe, Merv, Dalverzin-tepe, Ak-Beshim, Kuva, Krasnorechensk 
urban region, and many others 44 - over 100 inscriptions have been found, both on whole 
pottery vessels and on shards. The majority of these relate to the Kara-tepe monastery 

44 Staviskiy, 1963, pp. 171-6; Litvinsky, 1971, pp. 110-33. 



ISBN 978-92-3-103211-0 Contents Inscriptions in Buddhist complexes 

and have been published in the works of Grek, Vertogradova, Harmatta and Vorobyova- 
Desyatovskaya. 45 

An important historical and cultural inference can be drawn from a study of the inscrip- 
tions on ceramic ware and other articles: it is that followers of the Mahasamghika Hinayana 
school settled in the Old Termez region. This is the only written evidence to show that, in 
the Kushan period, the doctrine of this school permeated to the extreme westerly point of 
the Kushan Empire. It is known that Mahasamghika was one of the most ancient Buddhist 
schools which broke away after the schism within early Buddhism in c. 350 B.C. According 
to the written sources, the language in which this school started to codify its doctrine was 
Prakrit. The inscriptions found show that, in the west, this Prakrit was Gandhari, written in 
KharosthT script. 

Another Buddhist school which, judging by inscriptions, permeated beyond the fron- 
tiers of India was Dharmaguptaka. A third school, the name of which is attested in Bud- 
dhist manuscripts found in Central Asia, was Sarvastivada. Its establishment belongs to the 
period of Kanishka's rule, and evidently took place in Kashmir. The language employed 
by the adherents of the Sarvastivada school, and in which, according to Buddhist tradition, 
they codified their canon at the time of the Kashmir Council in the early second century, 
was Sanskrit. The findings of written sources have confirmed the information derived from 
Buddhist sources. The adherents of the Dharmaguptaka and Sarvastivada schools used the 
Brahmi script and the Sanskrit language. It is possible that the Kara-tepe ceramic inscrip- 
tions in Brahmi characters were made by followers of the Sarvastivada school. 

The written sources from Central Asia do not yet enable us to discuss the dissemina- 
tion of Mahayana in this area. The local inhabitants evidently continued their traditional 
adherence to Hinayana, which, to judge from the paintings and sculptures found in Bud- 
dhist temples and monasteries, broadly represented early Buddhism in all its aspects. The 
division of Buddhism into two main systems - Hinayana and Mahayana - took place in 
Indian territory only at the beginning of the Christian era. Starting from religious dogma 
(the sutras), it then made the transition to philosophical doctrine and the methods and func- 
tions of the mental techniques of yoga practice. The Buddhist texts found in manuscripts in 
East Turkestan have enabled us to follow how this division took place, and how, in the core 
of Hinduism and early Buddhism, the foundations of the three main systems (Hinayana, 
Mahayana and Vajrayana) emerged, systems which have persisted to the present day. 

The names of personal donors show that they included several Buddhist pilgrims from 
India, chiefly from the north-west. It is probable that the monasteries were founded by 

45 Grek, 1964, 1972; Vertogradova, 1983, p. 87; Harmatta, 1969, pp. 32-9; Vorobyova-Desyatovskaya, 



ISBN 978-92-3-103211-0 Contents Manuscripts from East Turkestan 

missionaries from that area, and the scribal schools that existed in the Kara-tepe monastery 
were linked to the scribal traditions of Kashmir and north-western India. The bulk of the 
donations were made by local inhabitants with Iranian names or names whose origin is 
unknown. People came to the monasteries with their relatives and probably lived there 
for some time. Among the donors were monks and lay people, and persons who performed 
economic functions within the monasteries. All of them made a 'gift of faith' (deya-dharma 
or dana-mukha) for the sake of their health and the health of those close to them, in order 
to ensure eternal life, to increase their religious merit and for favourable rebirths. 

Manuscripts from East Turkestan and Central Asia 

The manuscripts found in the oases of East Turkestan, Gilgit and Central Asia represent the 
canonical literature of a number of Hinayana and Mahay ana schools, together with tantras 
and a quantity of writings of the Vajrayana school, which was established in Central Asia 
at the end of the first millennium. The establishment of Buddhist doctrines in this area took 
place simultaneously in a variety of cultural traditions, between which there were close 
contacts and much interaction. Without recourse to Chinese and Tibetan translations of 
Buddhist sutras, it is impossible to follow the history of Buddhist texts in other languages. 

Collections of Sanskrit Buddhist manuscripts, copied in East Turkestan, are at present 
preserved in Europe, Japan and India. Manuscripts found in Middle Asia are preserved 
in St Petersburg, Ashgabat, Tashkent, Bishkek and Delhi. The total number of Sanskrit 
manuscripts probably amounts to some 5,000 preserved items (around 100,000 sheets and 

The earliest manuscripts in Brahmi characters known to science date from the Kushan 
period. They comprise three manuscripts from the German Turfan collection, one manu- 
script from Bamiyan, numerous small manuscript fragments on palm leaves from the col- 
lections of Petrovsky and Berezovsky (probably of the second-third centuries) and some 
sheets of a manuscript found in Bairam-Ali (probably dating from the third-fourth cen- 

The manuscripts confirm that, during the Kushan period, the Buddhist doctrine had 
already acquired a fairly large number of adherents throughout the territories of East 
Turkestan and Central Asia, and that their contents encompass excerpts from different 
parts of the Buddhist canon and passages from various Buddhist authors. The Turfan 
manuscripts, for example, include excerpts from the writings of the Indian scholar Ash- 
vaghosha, one of the first authors of Buddhist drama and poetry. Indian tradition holds 
that Ashvaghosha was a contemporary of Kanishka, who was his mentor and patron. After 



ISBN 978-92-3-103211-0 Contents Manuscripts from East Turkestan 

Kanishka's death, Ashvaghosha set down his actions in poetry. The creative heritage of 
Ashvaghosha in Central Asia is represented by passages from all his main works: the 
poems known as the Buddhacarita, which describe the life of Buddha Shakyamuni; the 
poems entitled the Saundaranandakdvya; and the Buddhist drama the Sdriputraprakarana, 
a biography of Shariputra, one of the Buddha's pupils. Ashvaghosha 's drama was unknown 
to Indian literature prior to the discovery of the early manuscripts in East Turkestan. 

The presence of Buddhist literary texts in Sanskrit of the Kushan period goes hand in 
hand with the codification of the Sanskrit canon of the Sarvastivada school in Kashmir at 
the Buddhist council in the time of Kanishka. Although there is no doubt that sutras already 
existed in Sanskrit at this time, the question as to which type of Sanskrit was represented in 
these early sutras has not finally been resolved, despite much research. Investigation into 
the language of Buddhist texts started at the beginning of the twentieth century with the 
works of Liiders. The designation of this language as 'Hybrid Sanskrit' was introduced 
following the appearance in 1953 of Edgerton's grammar and dictionary. 46 In the view 
of modern scholars, the term 'hybrid language' arises from the 'transposition' of Prakrit 
texts into Sanskrit, as a transitional stage to literary Sanskrit. As investigations have shown, 
grammatical and orthographic peculiarities depend not only on the degree of grammatical 
competence and education of the communicator, but also on the special features of the 
mother tongue. It seems more neutral to refer to this Sanskrit as 'Buddhist', which also 
designates the sphere in which it was used. 

It is clear that, in the fifth century, Sanskrit canonical texts started to be copied on a 
regular basis in local Buddhist centres in East Turkestan, chiefly in Khotan. At the same 
time, the translation of Buddhist texts into the Khotanese Saka language was also under- 
taken in Khotan. No information is available concerning the starting point for translations 
of Buddhist texts into the Tokharian languages. The earliest Tokharian manuscripts, copied 
in the northern oases of East Turkestan, date from the seventh century. 

The gradual transformation of East Turkestan into an international Buddhist centre, 
in which great numbers of manuscripts were copied, translated into various languages, 
transmitted to both East and West and preserved in monastic libraries and the houses of 
local inhabitants, is attested by the discovery here of a large number of copies of these 
texts in various languages and several scripts. The earliest Tibetan Buddhist manuscripts 
(dating from the mideighth century) have been found in East Turkestan, and it was here 
that the Tibetans began to translate Mahayana canonical texts into their language. 

In the northern oases, Tokharian, Uighur and Chinese translators worked on 
translations of Buddhist texts throughout the second half of the first millennium. The 

46 Edgerton, 1953. 



ISBN 978-92-3-103211-0 Contents Manuscripts from East Turkestan 

monasteries of Dunhuang maintained close cultural contacts with Turfan and Kucha. The 
earliest Chinese manuscripts from the Dunhuang library belong to the third-fourth cen- 
turies. Buddhist translation and literary activity in the Turkic languages started at a fairly 
early date but, with a few exceptions, the written texts of the seventh-ninth centuries are 
modest. The colophons of some manuscripts also mention translations from Tokharian A 
('the language of Karashahr') and Tokharian B ('the language of Kucha'). A number of 
translations from Sanskrit and (probably) Sogdian are known. 

It is important to note that the majority of Sanskrit Buddhist texts found in East Turkestan 
are not complete works, but merely excerpts and fragments, without headings or colophons. 
Due, however, to the unbroken tradition, from the manuscripts found in East Turkestan and 
Dunhuang, the Buddhist Hinayana canon and the principal Mahayana schools are the most 
fully represented in translations. This has allowed researchers to assess the structure of 
the canons of the various Hinayana schools, and the scope of the canonical literature of 


In the first millennium, the Hinayana, Mahayana and Vajrayana systems of Buddhism were 
represented in the region. No details are known about the sharp contradictions and hostili- 
ties that existed between the advocates of these systems, although it seems that the masses 
followed those who were victorious in open dispute. Between the northern and southern 
oases, however, there was a traditional doctrinal divergence connected with the cultural 
milieux defined by the Tokharian and Saka languages. Basically, Hinayana texts were pop- 
ular in Turfan, Kucha and Karashahr and a huge body of Buddhist Hinayana literature was 
preserved in translation. Malov has suggested that virtually the entire Hinayana canon was 
translated by the Uighurs into their language and kept in monastic libraries. 47 

In the southern oases - Khotan and Kashgar - the findings have mainly been of 
Mahayana texts since only the texts of this school were translated into Khotanese and 
Tibetan. The great majority of Chinese manuscripts from Dunhuang contain Mahayana 
writings, with an insignificant number of Hinayana texts. In Middle Asia, on the other 
hand, only Hinayana writings have so far been found and we have no information concern- 
ing the dissemination of Mahayana in the region. 

Towards the end of the first millennium, Vajrayana writings, sutras with a large number 
of incantational dhdrams (formulae), collections of dhdranis, magical formulized mantras, 
and so on, appeared throughout the territory of East Turkestan: they were in the Sanskrit, 

47 Malov, 1951, p. 142; Laufer, 1907, p. 302. 



ISBN 978-92-3-103211-0 


Manuscripts from East Turkestan 

FIG. 5. Wooden tablets with incantational dharani. Script in North Turkestan Brahmi. Found in 
Kucha by the archaeological mission of M. M. Berezovskiy. 

Tibetan, Khotanese, Chinese and Uighur languages (Fig. 5). A rupture in the fundamental 
doctrines of Buddhism is apparent from this material. Mahayana, with its basic concept of 
Nirvana and its attainment by the path of the 'great wisdom' (mahaprajna), gave way to 
more attractive and quicker methods, realizable in a single life-span, of achieving salvation 
in ' Sukhavati, the world of bliss', the 'pure land of Bodhisattva'. In difficult situations 
in life, recourse was had to the reading of dharams and mantras, which explains their 
great popularity. Many magical texts were inserted into the Mahayana texts, thus achiev- 
ing canonical status. As a result, Mahayana sutras began to be used for the attainment of 
practical blessings in life. 

Among the Hinayana sutras, those which included jatakas and avadanas (stories in 
which Buddhas or Bodhisattvas were reborn in human or animal form) continued to spread. 
Moral and ethical exhortations in verse form (of the Dharmapada type) were exceptionally 
popular. These not infrequently served as models for poetic imitations: Buddhist poetry in 
the Turkic languages, for example, includes 1,500 verses of this kind, in which the cus- 
tomary and favourite forms of folk poetry were used. Collections of jatakas and avadanas 



ISBN 978-92-3-103211-0 Contents Manuscripts from East Turkestan 

appeared that had been borrowed from the Buddhist canon. A synopsis of one such col- 
lection, compiled for Buddhist homilies, was found in Bairam-Ali in manuscripts of the 
third-fourth centuries. Analogous collections have been preserved in Chinese and Tibetan 

The pragmatic character assumed by Buddhist doctrine towards the end of the first 
millennium was evidence of its incipient decline. With the loss of political backing by 
local rulers, Buddhism was forced to relinquish its status as an independent religion and 
make the transition to the sphere of spells and magic cults. 


A second conclusion that can be drawn from the manuscripts found in East Turkestan 
and Central Asia is the existence of Sanskrit canons for most of the Hinayana schools. 
The 1982 Gottingen symposium established the existence of at least seven different ver- 
sions of the Sutrapitaka, relating to different schools. Fragments have been found of the 
Vinayapitaka of three schools, represented by the adherents of Sarvastivada, Mulasarvas- 
tivada and Dharmaguptaka. Research has shown that the majority of the Hinayana texts 
from East Turkestan belong to the Sarvastivada and Mulasarvastivada schools. 

The structure of the Hinayana canons, and the distribution of the sutras over the various 
dgamas, (sections), has not yet been precisely established. Sutras have been found which 
relate to four dgamas: DTrghdgama, Madhyamdgama, Samyuktagama and Anguttardgama. 
The proper arrangement of the dgamas within canons of the various schools has not been 
settled. A number of sutras were particularly popular, and were used independently of 
their relationships to the dgamas. The finding in Bairam-Ali of a manuscript in which the 
colophon has been preserved enabled the structure of the Vinayapitaka of the adherents of 
Sarvastivada to be determined precisely (Fig. 6). Besides the general compositions for all 
schools (the Prdtimoksa sutra and comments on it), the canon was found to include special 
writings of the karma-vdcana type, i.e. collections of rules for the performance of rites by 
monks and nuns. 


The range of Mahayana writings discovered in manuscripts in East Turkestan does not 
enable us to assert that this school of Buddhism possessed its own canon in India, con- 
structed on the same principle as the Hinayana canon, the Tripitaka. In Sanskrit, as in the 
other languages, only sutras have been preserved. Fragments of 25 Mahayana sutras have, 



ISBN 978-92-3-103211-0 Contents Manuscripts from East Turkestan 

FIG. 6. Manuscript on birch bark from Bairam-Ali (Merv oasis). 
Script in Indian BrahmT. Text from Buddhist canon of Sarvastivada. 

for example, been preserved in the Khotanese Saka language. 48 Khotan maintained per- 
manent links with Kashmir and the monasteries of northern India. The range of literature 
in circulation in Khotan has been found to be identical to that discovered in the Gilgit 
library. 49 

Central Asia has yielded a large quantity of Mahayana texts, allowing a picture to be 
drawn up of the Buddhist schools and cults that held sway in the various cultural tradi- 
tions. Among the schools, that of the Prajndpdramitd held the foremost position. Signifi- 
cant quantities of its manuscripts have been found in all languages (except the Tokharian 
languages). The Prajndpdramitd literature has been well researched, thanks to the work of 
Conze. 50 

The cult of the Buddha Amitabha, the lord of the west in Sukhavati (the world of bliss) 
was particularly popular among the peoples of Central Asia. In East Turkestan, the cult 
underwent further development and in the second half of the eighth century a special, 
apocryphal version known as the Aparamitdyur sutra arose, which was disseminated in all 
the local languages. The manuscripts found in Gilgit show that in India, too, the doctrine 
of ' Sukhavati, the world of bliss', became, in the second half of the first millennium, one 
of the chief Mahayana concepts. 

In East Turkestan, yet another Mahayana school - that of the Saddharmapundarikd, or 
'Lotus sutra' - experienced an important transformation. From this, the cult of the Bod- 
hisattva Avalokiteshvara separated out and underwent further, independent development. 
In East Turkestan, the 25th chapter of the 'Lotus sutra' was translated into every language, 
and the Bodhisattva Avalokiteshvara, who was eulogized in it, was transformed into a 

48 Emmerick, 1992. 

49 Hiniiber, 1979. 

50 Conze, 1977. 



ISBN 978-92-3-10321 1-0 Contents East Turkestan - a major Buddhist centre 

major deity, providing a shield against all earthly ills and gaining the status of protector of 

The sutra of 'Golden Radiance' - the Suvarnaprabhdsa sutra - underwent an analogous 
development in Central Asia. In the transformation of its text, preserved in a variety of ver- 
sions and editions in many languages, it is possible to follow the growth of basic Mahayana 
doctrine into Vajrayana. The dhdrams from the sutra of 'Golden Radiance' were also 
disseminated as a means of protection, and the reading of its text - thanks to a special pref- 
ace added to it in East Turkestan that was in harmony with Buddhist tradition - afforded 
protection against being reborn in hell. 

Besides the sutras existing in the popular milieu, important philosophical Mahayana 
sutras have been found in East Turkestan. These bear witness to the dissemination of the 
doctrines of two philosophical schools: Madhyamaka, which was established in India by 
Nagarjuna (150-250), and Yogacara, which derived its tradition from Asanga (310-390). 
An almost complete Sanskrit text of the sutra which provides the fundamentals of both 
schools (the Mahdratnakuta-dharmaparyaya or Kdsyapaparivarta sutra) has been discov- 
ered in East Turkestan. 

East Turkestan - a major Buddhist centre 

The pinpointing by archaeologists of the sites where manuscripts have been found, their 
palaeography, and the linguistic peculiarities preserved in their texts show that Buddhist 
centres were active in the territory of East Turkestan in the first millennium. In these cen- 
tres (Turfan in the north and Khotan in the south), Buddhist texts were translated from 
Sanskrit and then Prakrit languages into the local idioms, and were then disseminated by 
copying and printing from wooden blocks (xylography). The centres' activities encom- 
passed areas in which two local cultural traditions - the Tokharian and the Khotanese - 
were diffused. Both Turfan and Khotan experienced the powerful influence of the Indian 
and Iranian cultures, while in the north the cultural influence of China, and in the south 
that of Tibet, was felt. 

The high level of literacy of the population of East Turkestan and Dunhuang was 
undoubtedly linked to the presence of a large number of Buddhist monasteries. In Tokhar- 
ian manuscripts, the most frequently mentioned monastery is that of Yurpishka in Shorchuk, 
in the Kucha district, which was a major centre for the copying of manuscripts during the 
period of Hsiiantsang's visit to Kucha in the first half of the seventh century. From 650, 
the role of Karashahr probably started to grow (Kucha was devastated by the Chinese and 
the Turks) and in the eighth-ninth centuries Turfan assumed the leading position. In the 



ISBN 978-92-3-10321 1-0 Contents East Turkestan - a major Buddhist centre 

middle of the ninth century, after the disintegration of the Uighur Kaghanate, one Uighur 
branch established a small state in the Turfan oasis with its centre at Beshbalyk - later to 
be known as the Kocho state. The Buddhist centre moved there and, by the middle of the 
thirteenth century, Kocho was the single small island of Buddhist culture in the ocean of 
Islam, which engulfed the whole territory of East Turkestan. 51 

In the south, at the end of the seventh century, Khotan was gradually becoming an 
international Buddhist centre, encouraged by the tense political situation in the neighbour- 
ing countries. A period of wars and internal disorders had radically altered the situation in 
Central Asia and India. Among the inauspicious events for Buddhism, the following should 
be mentioned: the seizure of Central Asia by the Arabs in the eighth-ninth centuries; the 
persecution of Buddhism in T'ang China; the attempt to eradicate Buddhism from Tibet 
during the reign of gLang-dar-ma (836-842); the decline of Buddhism in India and the 
schism within the Buddhist community there; and, finally, the Muslim conquest of India. 
It is evident from the documents that regal power in Khotan continued to be strong, and 
afforded protection to Buddhism. The increase in the number of Buddhists is attested by 
the growing number of Buddhist centres referred to in the written sources in Tibetan, and 
of copies of Buddhist writings dating from the eighth and ninth centuries. 

It is probable that this period saw an influx into Khotan of Buddhists from neighbouring 
regions. Vajrayana texts circulated in Khotan at this time and Tibetans, too, participated 
in the creation of Vajrayana cults. It was perhaps in Khotan that the features of Tibetan 
tantrism developed that were to determine the fortunes of Buddhism in Tibet. The active 
and creative role of the Khotanese Buddhist centres is clear from the part they played in 
formulating the Buddhist canon and perfecting its structure. For example, at least three 
collections of sutras were compiled in East Turkestan that were not in circulation in India: 
the Maharatnakuta, the Mahavatamsaka and the Mahdsannipata. The assimilation of Bud- 
dhist doctrine and Indian culture evidently proceeded more vigorously in Khotan than in 
the northern oases. 

Translations from Sanskrit and the Prakrit languages, and the compilation of collec- 
tions of sutras and didactic, narrative writings, took place not only in Khotanese Saka, but 
also in the Tibetan and Chinese languages. It was in Khotan that the earlier, unaltered ver- 
sions of many Mahayana sutras were preserved, and Chinese pilgrims in quest of Buddhist 
works started to travel not to India, but to Khotan. Finally, Khotan was the site of the cre- 
ation of apocryphal works, on the basis of which the cults of individual Buddhas, patron 

51 Gabain, 1973. In 1986, Hamilton published 36 Uighur fragments of the ninth and tenth centuries from 
the P. Pelliot Collection and the British Library. These include many unidentified Buddhist fragments. The 
manuscripts relate to the period of existence of the Kocho state (see Hamilton, 1986). 



ISBN 978-92-3-10321 1-0 Contents East Turkestan - a major Buddhist centre 

Bodhisattvas and protectors evolved, as well as a number of abstract Buddhist categories 
of evil personified as demons and evil spirits, such as the kleshas. 

The foregoing justifies the assertion that, towards the end of the first millennium, Khotan 
was transformed into an international Buddhist centre of the East. This was the region in 
which Sanskrit versions of the Buddhist sutras were in circulation; these versions were 
then adopted as the basis of Chinese, Tibetan, Tangut, Uighur and Mongol Buddhism. 

The Buddhist centre in Khotan ceased to exist at the beginning of the eleventh century 
after its seizure by the Turks. As mentioned above, the sole Buddhist centre that continued 
to function in this region was Kocho, where the basic translations of Buddhist literature into 
the Uighur language were codified in written form. However, this lies outside the period 
covered by the present volume. 52 

52 We express our deep gratitude to S. G. Klyashtorny for giving us access to his work before it was 
published (Klyashtorny, 1992). See also Litvinsky, 1992. 



ISBN 978-92-3-103211-0 Contents THE ARAB CONQUEST 


B. A. Litvinsky, A. H. Jalilov and A. I. Kolesnikov 



The first Arab invasion of Iran 444 

The conquest of Iraq 445 

The conquest of Khuzistan 447 

The battle of Nihavend 448 

The conquest of Seistan 449 

The conquest of Khurasan 450 


The first Arab incursions into Transoxania 452 

The beginning of the conquest of Transoxania 453 

The struggle of the peoples of Central Asia against the Umayyads 454 

The struggle of the peoples of Central Asia against the c Abbasids 458 


Reasons for the fall of the Sasanian Empire 461 

Relations between Muslims and non-Muslims in the first century A.H 462 

First steps towards Islamization 463 

Iranian regional administration in the conquered territories 464 

Tabaristan and Dailam 465 

Zabul, Kabul, Gandhara and Ghur 466 

The Arabs in Sind 467 

The survival of pre-Islamic civilization 467 

See Map 5. 



ISBN 978-92-3-10321 1-0 Contents The first Arab invasion of Iran 

Part One 

(B. A. Litvinsky) 

The first Arab invasion of Iran 

A new religion - Islam - was founded by Muhammad (d. 632) in Arabia at the beginning 
of the seventh century. In the ensuing conquests, the Arabs subjugated the peoples of the 
Near and Middle East and of North Africa, and a vast territory - extending from Spain to 
Sind at the time of maximum expansion and including the western part of Central Asia 
- came under Arab rule. The historical destiny of the peoples of Central Asia was to be 
profoundly influenced by the Arab conquest and the spread of Islam. 

The Arabs advanced into Central Asia through Iran and so it is with Iran that the present 
account begins. 1 Arab tribes had settled in Mesopotamia even before the Sasanian era and 
the Sasanian Empire was therefore obliged to have dealings with them from the outset. 
According to al-Tabari, the Sasanian emperor Shapur I (241-271) actually settled one of 
the Arab tribes within Iran in Kerman. 2 Shapur II (309-379) subjugated the entire western 
part of the Persian Gulf. Islands were also incorporated in the Sasanian Empire and the 
Arab sea trade, linking the Mediterranean with India, was controlled by the Sasanians. 
Moreover, under Khusrau I (53 1-579) Iran intervened in the affairs of Yemen in an attempt 
to assist the Arabs against Byzantium. 

Sasanian Iran also controlled the semi-independent Arab kingdom of al-Hira (located to 
the north of the great swamp of lower Iraq) under the Lakhmid dynasty. Under the reign of 
Nu c man III (580-602), al-Hira had become increasingly independent. In spite of Nu c man's 
attempts to pursue an independent policy that was inconsistent with or even contrary to 

1 This account mainly follows the ideas and materials set out in the monograph written on the Arab 
conquest of Iran by Kolesnikov, 1982. 

2 Noldeke, Tabari, 1973, pp. 16-17. 



ISBN 978-92-3-10321 1-0 Contents The conquest of Iraq 

Sasanian interests, however, the small kingdom was incapable of holding its own against 
its giant neighbour and was subdued. Al-Hira and its neighbouring Arab tribes nevertheless 
continued to be a thorn in the side of the Sasanian authorities (in 604- 605, for example, 
they inflicted a defeat on the Sasanian forces). 3 Raids by individual Arab tribes inside the 
boundaries of the Sasanian Empire also continued. 

In the 630s, the 16-year-old shahanshah Yazdgird III (632-651) came to power in Iran; 
a group of dignitaries acted as his guardians and a regent governed on his behalf. The rule of 
the first caliph, Abu Bakr (632-634), began at almost exactly the same time. In the spring of 
633 the caliphate's forces under Khalid b. al-Walid embarked on their first campaign within 
Sasanian territory. They were opposed by the troops of one of the most powerful Iranian 
grandees, the governor of the border zone. The Iranian soldiers were fastened together by 
a chain to prevent them from fleeing. In the preliminary single combat, Khalid b. al-Walid 
slew his Iranian adversary and the Arabs were victorious in the subsequent attack, gaining 
possession of the shahanshah's crown valued at 100,000 drachms. Whereas 'The Persians 
had probably regarded the Muslim advance on al-Hira as another annoying raid of the 
bedouins,' 4 the battle of Kadhima opened up the route to Iraq for the Arabs. 5 

The conquest of Iraq 

The Arabs then set out to conquer Iraq, one army moving northwards from the Tigris- 
Euphrates delta and another moving across from the west. At a place near the harbour 
of c Ubulla, the Arabs first joined battle with an Iranian army sent against them from the 
capital, Ctesiphon: the Arabs were again victorious. They next launched an attack against 
al-Hira, hitherto the main Sasanian stronghold west of the Euphrates and the key to the 
inner regions of Iraq. Although the Sasanian army was defeated, the local population con- 
tinued to offer resistance for a time before eventually surrendering and paying tribute. 

All these events occurred prior to the accession to the caliphate of c Umar b. al-Khattab 
(634-644), under whom the forces of Islam enjoyed further successes. During his rule 
the Arab army in Iraq came under a new military leader, Abu c Ubaida. The Iranians also 
appointed a new commander-in-chief, the governor of Khurasan, Rustam b. Farrukhzad. 
Armour-clad war elephants helped the Iranians win one of the battles and Abu c Ubaida 
himself was crushed to death by an elephant. 6 In subsequent battles, however, the Arabs 

3 Kolesnikov, 1970, pp. 74-6. 

4 Frye, 1975, p. 56. 

5 Al-Tabari, 1879-89, Vol. 1, pp. 2023-5; Ibn al-Athir, 1851-76, Vol. 2, p. 295. 

6 Al-Tabari, 1879-89, Vol. 1, pp. 2175-6; Ibn al-Athir, 1851-76, Vol. 2, p. 388; al-Dinawari, 1888, p. 119; 
al-Baladhuri, 1866, p. 404. 



ISBN 978-92-3-10321 1-0 Contents The conquest of Iraq 

regained the upper hand and their territorial expansion continued. Their victory in the sec- 
ond battle of Yarmuk in August 636 completed the conquest of Syria and the caliph sent 
part of the Syrian army to the Mesopotamian front, their ranks swelled by large numbers 
of fresh volunteers. c Umar himself had intended to command these forces, but he subse- 
quently assigned the task to one of the Prophet's first companions, Sa c d b. Abu Waqqas. 
Altogether, the Arab troops in Mesopotamia numbered some 30,000, the largest Muslim 
military force ever assembled in that country. 

Sa c d b. Abu Waqqas then sent an embassy of 14 men to the Iranian capital; they 
demanded land concessions beyond the Euphrates, a trading corridor and the right to trade 
in Mesopotamia, together with the payment of taxes and tribute. The shahanshah rejected 
these demands out of hand. Arab sources give grossly exaggerated accounts of the size 
of the Iranian army that was subsequently raised. These estimates range from 60,000 to 
120,000, some sources even maintaining that the figure of 60,000 refers only to the so- 
called 'professional' soldiers of the regular army, who were accompanied by 'assistants 
and slaves'. According to modern researchers, however, the Iranian forces probably did 
not greatly outnumber the Arabs. 


The Iranians moved southwards until they were close to al-Qadisiyya, a small fortified town 
some 30 km from Kufa, and subsequent events took place on the bank of the c Atiq channel. 
To begin with, the Iranian general, Rustam b. Farrukhzad, again engaged in negotiations 
that were as protracted as they were futile. Then his men forced their way over the channel 
by a specially constructed crossing. They were drawn up in line along the channel, with 
the centre reinforced by a group of 18 war elephants, and smaller numbers defending the 
2 flanks. The elephants' attack proved highly effective and the Arabs were terrified. As a 
result, the Iranians initially had the upper hand, but the Arabs soon recovered and began 
to strike the elephants' trunks with their spears while the archers shot down their Persian 
counterparts. The Iranians were forced to retreat to their original positions. 

On the second day, the Arabs were more frequently on the offensive. The Iranians' war 
elephants were severely wounded and their opponents mounted palanquin-like structures 
on camels, causing panic among the Iranian horses. The Arabs were also reinforced with 
a 10,000-strong detachment from Syria. On the third day, both sides fought even more 
doggedly. The Iranians introduced new war elephants but without success: the wounded 
animals merely retreated into their own lines, sowing panic and confusion. The fighting was 
so fierce that it continued even after nightfall. Although the Arabs succeeded in breaking 
through the centre of the Iranian lines, both flanks stood firm. Nature itself was against the 



ISBN 978-92-3-10321 1-0 Contents The conquest of Khuzistan 

Iranians, who were lashed by winds of hurricane force. One of the Arab units fought its 
way through to Rustam b. Farrukhzad's headquarters and the general was killed in flight 
(according to another version, he died in single combat with Sa c d b. Abu Waqqas). 

The Iranian troops retreated in panic to the eastern bank of the channel. Although 
individual contingents continued to offer staunch resistance, the main group was anni- 
hilated and the rest fell back. There were extremely heavy casualties. The Arabs seized 
the Sasanian imperial flag ornamented with precious stones. 7 The sources disagree on the 
date of the battle of al-Qadisiyya; according to Kolesnikov, 8 it took place at the end of 
September 636. 

The Arab forces proceeded to capture the Iranian strongholds one by one, gradually 
drawing nearer to Ctesiphon. Yazdgird III himself fled the capital with his relatives and 
entourage and it surrendered without putting up any resistance. An enormously rich booty 
fell to the Arabs, including some of the contents of the state treasury that the fugitives had 
left behind in their haste. By 637 the whole of Mesopotamia had come under Arab control. 
According to Zarrinkub: 

On entering the palace of Khusrau, Sa e d had performed an eight rak c at prayer for his victory 
and, because of its appropriateness in recalling the fate of those who reject God, recited the 
Qur c anic verse (44, 25-27) which begins with 'How many gardens and springs have they 
left'. He made a mosque in the citadel and the four-hundredyear- old capital of the Sasanians 
became for a time the camping ground of this Muslim general. 9 

The conquest of Khuzistan 

The conquest of Khuzistan then began. The Arabs were led by Abu Musa Ash c ari, the 
future governor of Basra. He first captured two border strongholds, the outcome being 
determined by victory on the battlefield. In the meantime, Yazdgird had moved to Nihavend 
in central Media. One after another, the towns and strongholds of Khuzistan fell to the 
Arabs, including the provincial capital of Ahvaz. Then began the battle for Shushtar, which 
was of great strategic importance. A seventh-century Syriac chronicle describes the town 
as follows: 'Shushtar covers a wide area and is strongly fortified by large and deep chan- 
nels, surrounding it on all sides like moats.' The sources disagree on the duration of the 
battle, mentioning periods ranging from a few months to two years. The action began 
on the outer defence lines, where the Iranian troops occupied trenches. Having crushed 
the outer defences (the defenders retreated into the town, reinforcing its own defences), 

7 Yusuf, 1945. 

8 Kolesnikov, 1982. 

9 Zarrinkub, 1975, pp. 12-13. 



ISBN 978-92-3-10321 1-0 Contents The battle of Nihavend 

the Arabs stormed the fortifications but their attacks were repulsed. However, they were 
assisted by a traitor called Siya, who led an Arab detachment through a secret entrance 
by night. They killed the guards and threw open the gates. The remainder of the garrison 
sought refuge in the citadel, where the treasure was deposited. Their commander was Hur- 
muzan (or Hurmuzdan), the marzbdn (governor) of Khuzistan. Although the garrison put 
up fierce resistance, beating off the Arab attacks, the defenders' strength dwindled and the 
citadel eventually fell. According to some sources, the marzbdn sued for peace and his life 
was spared following his conversion to Islam; the majority of his companions, however, 
were put to death. 10 

The battle of Nihavend 

Finally realizing what a formidable adversary he was dealing with, Yazdgird III issued a 
decree for the mobilization of troops, especially from the neighbouring provinces. They 
were to assemble in Nihavend. The forces that gathered there came from Media itself, 
from Persia, the Caspian provinces, Khurasan, Seistan (modern Sistan) and other regions. 
A huge army was formed, estimated by the sources to number between 60,000 and 150,000 
men. The caliph c Umar also mobilized an impressive number of troops, some of whom 
came from Syria; Nu c man b. c Amr b. Muqarrin was appointed commander-in-chief. 

The Iranians drew up their troops close to Nihavend, firmly resolved to fight a defensive 
battle. Iron spikes were strewn in front of the lines to hinder the attacks of the Arab cavalry. 
Once again, the Iranian infantry were chained together in groups of five to ten men to 
prevent them retreating. After initial fruitless parleying, the Arabs spread the rumour that 
the caliph was dead and they were about to retreat, even going through some of the motions 
of withdrawal. The ruse worked and part of the Iranian army advanced into the open field. 
During the fierce fighting that resulted, both sides incurred heavy losses, but the scales 
eventually tipped in the Arabs' favour. Although their commanderin- chief was killed, the 
Arabs continued to attack and the Iranian army was routed, part of it seeking refuge in the 
town and the neighbouring fortress. The Iranians then abandoned the fortress, emerged on 
to the battlefield and continued fighting. They were utterly defeated. The ruler of Nihavend 
managed to conclude a charter of immunity for the population of the town, but the Arabs 
seized great quantities of booty, including the treasures of the fire temple. It is believed that 
these events occurred in 642. n 

10 Al-Kufi, 1968-75, Vol. 2, pp. 23-5; al-Baladhuri, 1866, pp. 380-1; al-Tabari, 1879-89, Vol. 1, pp. 
2554-5; Qummi, 1934, p. 300. 

11 Kolesnikov, 1982, p. 111. 



ISBN 978-92-3-10321 1-0 Contents The conquest of Seistan 

According to Frye, 'this was the most difficult battle of all those which the Arabs had to 
fight against imperial Sasanian forces'. 12 Described as the 'battle of battles', its role was 
extremely important since it took place on the Iranian plateau. After the battle of Nihavend, 
the organized resistance of the Sasanian authorities came to an end. According to al-Tabari, 
'from that day on, there was no further unity among them [the Persians] and the people of 
the individual provinces fought their own enemies on their own territory'. After the fall of 
Nihavend, Yazdgird III is reported to have moved to Istakhr and from there to Kerman and 
finally Khurasan. Following their victory in Nihavend, the Arabs captured Hamadan, which 
had to be subdued on two occasions. The whole of Media was now under their control. 

The sources do not give a clear-cut chronological account of the remainder of the con- 
quest. This confusion is frequently due to the fact that many towns and even provinces had 
to be conquered two or more times. Isfahan was taken in 643 and 644, for example, and 
Ray was captured at around the same time. The Arabs' next step was to subjugate Iran's 
northern provinces. They also took possession of Persia and Kerman. 

The conquest of Seistan 

We shall now consider in greater detail the Arab conquest of Seistan and Khurasan, regions 
that form part of Central Asia. Seistan (in ancient times, Drangiana) was one of the remote 
eastern provinces of the Sasanian Empire. The Arab conquest of Seistan began in the mid- 
dle period of the caliph c Uthman's rule (644-656), although the first raids had taken place 
under the previous caliph. c Uthman appointed his fellow tribesman, c Abdallah b. Amir, 
future governor of Basra, with instructions to complete the conquest. Having fortified his 
position in Kerman, c Abdallah b. Amir planned to advance on Seistan. An attack was possi- 
ble from the west, but this would have meant crossing the vast desert of Dasht-i Lut, which 
extended for a distance of some 450 km. The other approach was from the north-west via 
Kuhistan and Herat: it was a much easier route but those regions would first have to be 

It was nevertheless decided to take the route through the desert and the first offensive 
took place in 650-651. c Abdallah b. Amir placed the attacking forces under the command 
of Mujashi b. Mas c ud. Many Arabs were slain in the fighting and their forces had to retreat. 
The following year, c Abdallah b. Amir himself took part in the campaign: he proceeded 
with the main body of the army to Khurasan, while Rabi c b. Ziyad was sent to Seistan, 
where he succeeded in reaching the town of Zaliq, some 30 km from the capital Zarang. 
The ruler of Zaliq preferred a peace treaty to the battlefield. The Arabs then subjugated the 

12 Frye, 1975, p. 60. 



ISBN 978-92-3-10321 1-0 Contents The conquest of Khurasan 

towns of Karkuya, Haisun and Nashrudh before arriving in the vicinity of Zarang. They 
crossed the Huk, a channel or tributary of the River Helmand, and came to the walls of 
the capital. A fierce battle ensued and during the first attack many Muslims were killed. 
The second attack resulted in an Arab victory. But the town's fortifications, especially its 
citadel, were exceptionally strong - according to local tradition, they had been constructed 
by order of Kai Khusrau, Alexander the Great and Ardashir I Papakan, the cofounder of 
the Sasanian Empire. 

Iran b. Rustam, the marzban or ispahbad (local governor) of Zarang, then assembled 
the elite of the nobility and the Zoroastrian clergy, who agreed that hostilities should end 
even if the military commanders wished to continue fighting. A 'justification' was even 
found: the 'divine mission' of the Arab aliens was supposed to have been foretold in the 
Zoroastrian sacred books. Rabi c b. Ziyad agreed to make peace on condition that Seistan 
paid an annual tribute of 1 million dirhams, to be handed over by 1,000 boy slaves (other 
sources say girls), each bearing a golden bowl. The entire region of Seistan was then sub- 
jugated and the conqueror, Rabi c b. Ziyad, was appointed Arab governor. Arab- Sasanian 
coins minted in Seistan in 651-652 and 652-653 are known. When Rabi c b. Ziyad was 
recalled to Basra some 18 months later, he was replaced by Abu Sa c id c Abd al-Rahman 
b. Samura. In the meantime, however, the local inhabitants had risen up against the Arabs 
and overthrown them, and the new governor had to resume military action. Zarang was 
subdued and this time the conquest was final. The Arab forces had extended the territory 
under the caliph's rule as far as India. 13 

The conquest of Khurasan 

The conquest of Khurasan was bound up with the fate of the last Sasanian shahanshah. 
Abandoning Kerman, Yazdgird III had gone, according to most of the sources, straight to 
Khurasan (other versions say Seistan), halting in Nishapur (or, in another source, Bust). 
He then moved on to Merv because it was ruled by the kanarang (the east Iranian counter- 
part of marzban) Mahoe, who was personally indebted to Yazdgird for his ascent to high 
office. 14 The shahanshah apparently hoped to enlist the aid of the Turks and Chinese and 
raise a new army. 

In the meantime Yazdgird had no army. He travelled with a suite of several thousand 
relatives, courtiers and servants, accompanied (as far as Merv) by only a small military 
detachment. Friction then arose between Yazdgird and his vassal Mahoe. The sources offer 

13 Bosworth, 1968, pp. 13-25. 

14 For Khurasan on the eve of the Arab conquest, see Shaban, n.d. 



ISBN 978-92-3-10321 1-0 Contents The conquest of Khurasan 

different explanations for this conflict: some maintain that Mahoe had already come to an 
agreement with the Turk kaghan, hoping to secede from the Sasanian Empire; according to 
others, Yazdgird ordered Mahoe to pay a large tax; a third group claims that the shahanshah 
wished to remove Mahoe and appoint a military leader in his place, and when this plan 
failed sought to replace Mahoe by his nephew. According to one account, the people of 
Merv had also been turned against Yazdgird and refused him entry to the town. The Turk 
kaghan, whatever his role in the affair, can hardly have been well-disposed towards the 

What is certain is that a conspiracy had been hatched against Yazdgird. Having no troops 
at his disposal, he secretly abandoned his residence and took flight, hiding in a mill on the 
River Murghab with a Christian miller. The fugitive's costly apparel, jewellery and perfor- 
mance of the Zoroastrian rites made it easy for his pursuers to track him down. According 
to some sources, the miller himself murdered his illustrious guest. The funeral was orga- 
nized by Merv's Christian community, who buried him in the garden of the Metropolitan of 
Merv. (According to another report, however, his body was taken to Istakhr.) The sources 
also give divergent accounts of Mahoe's fate: according to some, he sought refuge from the 
approaching Arabs with the Turk kaghan; another source claims that the people of Merv 
themselves delivered their marzban to an agonizing death. 

Yazdgird's death in 65 1 finally brought an end to the Sasanian Empire, but it did not halt 
the Arab advance. Under c Abdallah b. Amir, the Arabs captured Nishapur, routing Hep- 
hthalite forces from Herat province in the process. Following a siege of several months, 
Nishapur was finally betrayed by a member of the Iranian aristocracy and its citadel cap- 
tured. Towns such as Tus, Abivard, Nisa, Sarakhs, Herat and Merv then fell to forces com- 
manded by the Arab general Ahnaf b. Qais. Continuing eastwards, he reached and cap- 
tured Balkh. Khurasan was subdued but the local population persistently rebelled against 
the Arabs. A long struggle still lay ahead. 



ISBN 978-92-3-10321 1-0 Contents The first Arab incursions into Transoxania 

Part Two 


(A H. Jalilov) 

The first Arab incursions into Transoxania 

Under the Umayyads, the conquest began in the 680s of parts of the right bank of the Amu 
Darya (an area known to the Arabs as Md ward 'l-nahr. literally, 'that which is beyond 
the river', i.e. Transoxania). The forces came from Khurasan, where an Arab governor- 
ship had been set up with the town of Merv as its centre. At first the campaigns took the 
form of predatory raids. The first major raid into Transoxania was carried out by the gov- 
ernor of Khurasan, c Ubaidallah b. Ziyad. In 673 he crossed the Amu Darya and reached 
Bukhara, which at that time was ruled by the khatun (queen), the mother of the young 
Bukhar khudat (ruler of Bukhara), Tughshada. After the very first skirmish she made peace 
with c Ubaidallah b. Ziyad, who obtained a ransom from her and returned to Merv. 

In 676 the Arabs repeated their raid on Bukhara under the leadership of the new gover- 
nor of Khurasan, Sa c id b. c Uthman. The khatun made peace with him too and he went on to 
Samarkand, having taken 80 hostages. 15 All the attempts by Sa c id b. c Uthman to capture 
the town proved unsuccessful: he was forced to make peace with the inhabitants and to 
leave the territory of Transoxania, taking 50 Sogdian hostages with him. On his return to 
Medina he made slaves of them, but they killed him and then committed suicide, preferring 
death to slavery. 16 

The Arabs next raided Khwarizm, Khujand and Samarkand in 680 under the leadership 
of the new governor of Khurasan, Salm b. Ziyad. Their rulers also made peace with him 
and, after obtaining a ransom from them, he withdrew from Transoxania. Similar raids were 
conducted by the next-but-one governor, al- Muhallab b. Abi Sufra, and his successors up 
to the year 705. 17 In spite of these raids, the local rulers still did not realize the seriousness 

15 Narshakhi, 1897, p. 52. 

16 Al-Baladhuri, 1866, p. 412; Gibb, 1923, pp. 18-19. 

17 For these raids, see Istoriya Tajikskogo naroda, 1964, pp. 96-7. 



ISBN 978-92-3-103211-0 Contents The beginning of the conquest of Transoxania 

of the Arab threat. Instead of uniting to repulse the foreign invaders, some rulers even 
invited their enemies into their country so that, with their help, they could settle accounts 
with neighbouring rulers. 18 

The beginning of the conquest of Transoxania 

At the end of the seventh/beginning of the eighth century, Arab policy towards Central 
Asia underwent a fundamental change. The internecine strife among the Arabs subsided 
somewhat towards the end of the reign of the caliph c Abd al-Malik b. Marwan (685-705), 
and the Umayyads were able to begin the systematic conquest of Transoxania. 

In the year 705, the task of conquering Transoxania was entrusted to the governor of 
Khurasan, Qutaiba b. Muslim (705-715), who ushered in a new and decisive stage in the 
conquest of Transoxania by the Umayyads. Qutaiba skilfully exploited the internal quar- 
rels between the rulers of Central Asia. During his first campaign in Transoxania, his forces 
included the ruler of Balkh, and the Chagdn khuddt (ruler of Chaganiyan) invited Qutaiba 
into his country to participate in a joint struggle against the ruler of the neighbouring ter- 
ritories of Akharun and Shuman. 19 Such treason provided Qutaiba with the information 
he required about Transoxania and in 706 he undertook a bold campaign in the area of 

One of the closest towns to the Amu Darya in the Bukhara oasis was the small trading 
centre of Paikent. This was the first place in which Qutaiba encountered stubborn resis- 
tance from the population of Transoxania. The Arabs were forced to take Paikent twice, 
killing all of its defenders and razing the town to the ground. In 707 and 708 Qutaiba 
attempted to seize the oasis of Bukhara, but was vigorously repulsed by the combined 
forces of Bukhara, Sughd and the Turks, and returned to Merv. It was only after making 
peace with the Sogdian ikhshid (king), Tarkhun, and driving a wedge between the allies 
that Qutaiba managed to capture Bukhara in 709, and Shuman, Kish and Nakhshab the 
following year. Tarkhun's policy of compliance with the invaders greatly displeased the 
Sogdians: in 710 they dethroned him and elected Ghurak in his place. 20 This was a conve- 
nient pretext for Qutaiba to begin the conquest of Sughd and Khwarizm. (For the conquest 
of Khwarizm, see Chapter 9.) 

In 712 the Arab commander set out against Samarkand, after incorporating military 
detachments from Khwarizm and Bukhara in his own main forces. The Sogdians first gave 
battle to the Arabs at Arbinjan; this was followed by the siege of Samarkand, which lasted 

18 Al-Tabari, 1879-89, Vol. 2, p. 994. 

19 Al-Baladhuri, 1866, pp. 419-20. 

20 Al-Tabari, 1879-89, Vol. 2, pp. 1229-30. 



ISBN 978-92-3-10321 1-0 Contents The struggle of the peoples of Central Asia 

a month. Although the rulers of Chach (Tashkent) and Ferghana sent a small detachment to 
assist the people of Samarkand, the Arabs managed to destroy the force before it arrived. 21 
Exhausted by the month-long siege, the Sogdians had no option other than to surrender and 
to make peace on the worst possible terms. In 713 and 714, Qutaiba conducted two major 
campaigns against Chach and Ferghana 22 and almost reached the territory of Kashgar. 

The following year, Sulaiman (715-717) succeeded to the caliphate. Qutaiba b. Muslim, 
aware of Sulaiman's hostility to him, moved with his family to Ferghana in order to break 
away from the caliphate. But the Arab troops, wearied by the continuous bloody wars 
which had lasted for a decade, would no longer obey Qutaiba, and he and his family were 
killed. For several years after his death, there were no more Arab conquests in Central Asia. 
With the exception of a raid on Kashgar and the conquest of Dihistan (on the shores of the 
Caspian Sea), the Arabs launched no major campaigns to extend their dominions during 
the period 715-720, concentrating instead on consolidating their hold on the regions they 
had already conquered. 

From the first stages of the conquest, the nomadic Arab nobility had attempted to col- 
onize the conquered areas of Transoxania. This policy was widely pursued by the Arabs, 
particularly under Qutaiba b. Muslim, who consolidated his military victories by settling 
Arabs among the population and through them conducting large-scale propaganda cam- 
paigns on behalf of Islam. 23 

The struggle of the peoples of Central Asia against 
the Umayyads 

In order to attract the people to Islam, the Arabs initially offered certain privileges to con- 
verts as well as applying methods of coercion. Those who accepted Islam, for example, 
were exempted from payment of the jizya (poll-tax). But when mass conversions began 
and tax receipts declined, the governor of Khurasan, al-Jarrah b. c Abdallah al-Hakami 
(717-719), decreed that only converts who accepted circumcision and were acquainted 
with the Qur c an would be exempted from payment of the jizya. This gave rise to the large- 
scale anti-Umayyad movement of the Sogdians in the years 720-722. 

There were two stages in this movement. During the first stage (720- 721), the Sogdi- 
ans, with the aid of the Turks, destroyed the Samarkand garrison and expelled the Arabs 
from the town. All the attempts by the governor of Khurasan, Sa c id b. c Abd al- c Aziz b. 

21 Ibid., Vol. 2, pp. 1242-3. 

22 Ibid., Vol. 2, p. 1256. 

23 Narshakhi, 1897, p. 63. 



ISBN 978-92-3-10321 1-0 Contents The struggle of the peoples of Central Asia 

al-Hakam (720-721), to restore Arab power in Samarkand proved unsuccessful. In the 
autumn of 721 he was replaced as governor by Sa c id b. c Amr al-Harashi. 

On his arrival in Khurasan, al-Harashi organized a major punitive expedition against 
the Sogdians. However, their ikhshid, Ghurak, instead of leading the rising, tried to per- 
suade his subjects to offer allegiance to al-Harashi. 24 For these reasons, the anti-Umayyad 
movement among the Sogdians then entered a second stage (721-722), moving from an 
active to a passive struggle. Realizing that their forces were inadequate, the Sogdian rebels 
left their homeland and moved to regions which offered greater protection from their foes. 
The rebels from the western part of Sughd, led by Karzanj, the ruler of Pai (presentday 
Katta-kurgan), set out for Ferghana, whose king, at-Tar, promised them protection and 
refuge. The rebels from the eastern part of Sughd, led by Divashtich, the ruler of Pan- 
jikent, travelled east to the upper reaches of the Zerafshan. But at-Tar proved perfidious; 
when the Sogdians arrived he held them in Khujand and secretly informed al-Harashi of 
their whereabouts. 25 The Arab governor swiftly dispatched a large detachment and dealt 
brutally with the Sogdian emigrants. He also killed over 3,000 farmers in the Khujand 
neighbourhood because of their solidarity with the Sogdians. 26 

After brutally annihilating the group led by Karzanj, al-Harashi rapidly dispatched a 
detachment against the Sogdians under Divashtich; they were occupying the fort of Abar- 
gar (now known as 'the castle on Mount Mug'), located on the left bank of the Zerafshan, 
some 120 km to the east of Panjikent. 27 At the approach of the Arabs, the Sogdians sortied 
and gave battle to the enemy at a distance of 6-7 km from the fort in a gorge near the vil- 
lage of Kum. The Arabs won and laid siege to the castle. Realizing that further resistance 
was useless, Divashtich gave himself up to the Arabs, who then seized and pillaged the 
fort. In the autumn of the same year (722), al-Harashi had Divashtich killed on the road 
from Kish to Arbinjan. 28 The defeat of this second group of rebels sealed the fate of the 
anti-Umayyad movement among the Sogdians in the years 720-722. 29 

Although the Arabs dealt harshly with the movement of 720-722 and reestablished their 
authority over Sughd, the people of Transoxania continued their resistance, this time with 

24 Al-Tabari, 1879-89, Vol. 2, p. 1,439; Ibn al-Athir, 1851-76, Vol. 5, p. 78. 

25 Al-Tabari, 1879-89, Vol. 2, pp. 1440-2; Ibn al-Athir, 1851-76, Vol. 5, pp. 78-9. 

26 Al-Tabari, 1879-89, Vol. 2, pp. 1445-6; Ibn al-Athir, 1851-76, Vol. 5, p. 81. 

27 In 1933 various artifacts, including over 80 manuscripts, were discovered in the ruins of the fort that 
had once stood on Mount Mug (see Sogdiyskie documenty s gory Mug, 1962, Vol. 1; 1962, Vol. 2; 1963, 
Vol. 3). The excavation of ancient Panjikent, whose last ruler was Divashtich, was begun in 1947 and is still 
under way. The ruins of this ancient town, which ceased to exist after being conquered by the Arabs, are 
located to the south-east of the modern town of the same name (Jalilov and Negmatov, 1969; Belenitskiy and 
Raspopova, 1971; Isakov, 1982). 

28 Al-Tabari, 1879-89, Vol. 2, p. 1447. 

29 For a detailed account of this movement, see Jalilov, 1961, pp. 134-46. 



ISBN 978-92-3-10321 1-0 Contents The struggle of the peoples of Central Asia 

the aid of the Turks. In 728, in an attempt to reduce popular discontent and consolidate 
Arab power in Transoxania, the governor of Khurasan, Ashras b. c Abdallah al-Sulami, 
decreed that anyone accepting Islam would be exempt from the jizya. So many people 
responded by becoming 'Muslims' that there was hardly anyone left to pay the jizya. But 
the abandoning of the tax conflicted with the interests of both the Arabs and the local elite. 
Al-Sulami therefore revoked his decision in the same year and again began to levy the 
jizya on all non-Muslims and on Muslims who had not yet been circumcized and were not 
familiar with the Qur c an. This led to a major rebellion which extended to almost the whole 
of Transoxania. 

The oasis of Bukhara became the centre of the rising, attracting rebels from Sughd and 
the Turks, led by their kaghan. The Arabs were practically driven out of Transoxania by 
a broad popular rising in 728: only Samarkand and Dabusiyya remained in their hands, 
and that was due to the indecisiveness of the ikhshid, Ghurak. 30 Al-Sulami only managed 
to recapture Bukhara in the summer of 729, after several months of hard fighting. In the 
spring of 730 a new governor of Khurasan, Junaid b. c Abd al-Rahman al-Murri, arrived 
in Bukhara to assist al-Sulami. Their joint forces reached Samarkand with great difficulty 
and, after consolidating the garrison there, Junaid returned to Khurasan the same year. 

From that time the Arab position became more difficult in Transoxania and also in 
Khurasan itself. In 733-734 there was a drought and famine broke out. 31 One cause of 
the famine in Khurasan was the Arabs' loss of the Zerafshan valley, which supplied them 
with large quantities of grain. In 734 an antigovernment movement led by Harith b. Suraij 
broke out in Khurasan among the Arabs themselves, but it was rapidly crushed by the new 
governor, Asad b. c Abdallah. 

Taking advantage of the troubles and disturbances among the Arabs, the people of Tran- 
soxania intensified their struggle against the invaders in the years 736-737. In response, 
Asad b. c Abdallah transferred his capital from Merv to Balkh and in 737 led an expedition 
to Khuttal. Armed forces from Sughd and Chach and numbers of Turks arrived to sup- 
port the population of Khuttal. The Turk kaghan, Sulu, emerged as their overall leader and 
the first blow was struck against the Chagan khudat, who had previously supported the 
Arabs. 32 Asad b. c Abdallah fled, leaving behind his baggage train containing the plunder 
from Khuttal. Sulu pursued Asad and on the left bank of the Amu Darya split his forces into 
small detachments, which then took to looting the countryside. On hearing this news, Asad, 
who was preparing to abandon Balkh for Merv, rapidly went over to the attack and won a 

30 Istoriya Uzbekskoy SSR, 1955, p. 147. 

31 Al-Tabari, 1879-89, Vol. 2, p. 1563. 

32 Ibid., pp. 1600-1. 



ISBN 978-92-3-10321 1-0 Contents The struggle of the peoples of Central Asia 

resounding victory over the allies not far from Kharistan. With the subsequent appointment 
of Nasr b. Sayyar (738-748) as governor of Khurasan and the collapse of the anti-caliphate 
coalition, the Arabs succeeded in consolidating their position in Transoxania. 

Having taken part in the Arab conquests in the time of Qutaiba b. Muslim, Nasr b. Say- 
yar knew Transoxania well and he realized that it would be impossible to subdue the coun- 
try by military action alone. He therefore attempted to normalize relations with the local 
population by peaceful means. He introduced a fixed procedure for the levying of taxes 
and attempted to establish close relations with the local elite, even marrying the daughter 
of the Bukhdr khuddt. Through several such measures, Nasr and his comrades managed to 
win influential groups of the local elite over to their side and began the process whereby 
the Arab aristocracy merged with the local elite. Nevertheless Nasr failed to restore order 
in Transoxania and discontent with Umayyad policies continued to grow, not only among 
the people of Transoxania and the other countries conquered by the Arabs but also among 
the Arab population itself. 

The power of the Umayyads rested on the aristocratic elite and protected its interests 
alone; as a result, the broad masses of the Arab population were dissatisfied with Umayyad 
rule. This hostility was particularly strong in Khurasan and Transoxania. In those areas, not 
only the lower sections of the population (who had to meet obligations such as the kharaj, 
or land tax, and the jizya) but even the local aristocracy harboured resentment; although 
the aristocracy had established close relations with the conquering elite, they did not enjoy 
the same rights. 

This general discontent was skilfully exploited by the c Abbasids (the descendants of 
c Abbas, the uncle of the Prophet Muhammad), who in the 740s secretly began to conduct 
a vigorous propaganda campaign against the Umayyads. At a crucial stage, the c Abbasids 
sent Abu Muslim, a man loyal to them, to Khurasan as leader of the movement. He enjoyed 
great success in Khurasan and Transoxania in 747-748: when the rebellion was raised, peo- 
ple flocked to join him under the black banners of the c Abbasids. Although Nasr b. Sayyar 
vainly attempted to rally the Arabs, his forces were destroyed and the rebels seized Merv 
and then the whole of Khurasan. Taking advantage of his success, Abu Muslim occupied 
Damascus, the seat of the caliphate, in the year 750. The power of the Umayyad caliphs 
collapsed, to be followed by a new Arab dynasty, the c Abbasids, with its first capital at 
Kufa (and, subsequently, its permanent capital at Baghdad). In the battle at the Talas river 
(751), the Arabs defeated the Chinese. 



ISBN 978-92-3-10321 1-0 Contents The struggle of the peoples of Central Asia 

The struggle of the peoples of Central Asia against 
the c Abbasids and the local nobility 

The fall of the Umayyads and the rise to power of the c Abbasids did little to alter the 
wretched conditions of the mass of the population. Like their predecessors, the c Abbasids 
were jealous defenders of Arab dominion over other conquered countries. Not one of 
the promises made to the people by the c Abbasids and the leader of the movement, Abu 
Muslim (appointed governor of Khurasan after the victory over the Umayyads), was ful- 
filled. The people still laboured under a host of burdensome obligations, provoking popular 
risings from the earliest years of c Abbasid rule. 

In 750 a rebellion erupted in Bukhara, directed not only against the c Abbasids but also 
against the local aristocracy which had sided with them. The rising was led by Sharik b. 
Shaikh, who encouraged his followers by saying that they had not fought the Umayyads 
merely in order to submit to the c Abbasids. Abu Muslim dispatched a force of 10,000 
against them, led by Ziyad b. Salih, but in the 37 days' fighting in Bukhara that ensued, the 
rebels were victorious in every battle. 33 The Bukhdr khuddt, Qutaiba, then came to the aid 
of Ziyad b. Salih with a force of 10,000. With the help of these soldiers (together with the 
fact that the rebels were suffering from a severe shortage of food), Ziyad seized the town 
and dealt harshly with its population. Sharik b. Shaikh was killed in one of the battles. 
There was a similar rising in Samarkand, which was also brutally suppressed by the same 
Ziyad b. Salih. Although Abu Muslim fought the rebels and jealously defended c Abbasid 
power in Khurasan and Transoxania, the c Abbasid rulers did not trust him and they had 
him murdered in 755. 

The murder of Abu Muslim gave rise to a number of rebellions against the c Abbasids. 
Although Abu Muslim had not been a true popular leader, the people saw in him the man 
who had freed them from the Umayyad yoke and had promised to improve their lot. In 
755 a rebellion broke out in Nishapur which spread to almost the whole of Khurasan 
and Tabaristan: it was led by Sumbad, who declared himself a follower of Abu Muslim. 
Although this rising too was brutally repressed by the c Abbasids, it provided a powerful 
stimulus for a larger-scale rebellion which broke out in the territory of Transoxania during 
the 770s and was this time headed by a genuine popular leader, Hashim b. Hakim, known 
by the nickname al-Muqanna c . 

Al-Muqanna c was born in one of the villages of Merv. In the years of the anti-Umayyad 
campaigns of the c Abbasids, he had been one of Abu Muslim's military commanders. In 
776, aware that the people of Transoxania were hostile to the policies of the c Abbasids, 

33 Narshakhi, 1897, p. 82. 



ISBN 978-92-3-10321 1-0 Contents The struggle of the peoples of Central Asia 

al-Muqanna c dispatched his emissaries to call the people to rise in open revolt against the 
foreign yoke and the inequality of their property status. A number of towns and villages 
in the Zerafshan valley and the Kashka Darya immediately announced their readiness to 
support al-Muqanna c , the inhabitants of 60 villages rallying to him on a single day. 34 Con- 
vinced that the number of his supporters in Transoxania was growing quickly, al-Muqanna c 
travelled to Kish with 36 followers. By the time he arrived, the Kashka Darya valley and 
the villages around Bukhara were already in the hands of his followers, 'the people in white 
clothes' (al-Muqanna c 's followers were distinguished by their white clothes and banners, 
whereas the c Abbasid colour was black). 

The caliph, al-Mahdi (775-785), sent a large force under Jibra'il b. Yahya to crush the 
rising. On arrival in Bukhara, Jibra'il, along with the ruler, Hussain b. Mu c az, attacked 
the village of Narshakh (a rebel strongpoint in the Bukhara area) and took it after a four- 
month siege. Jibra'il killed two of the rebel leaders, Hakim b. Ahmad and his comrade- 
in-arms, Khashvi. A new battle flared up by the walls of the settlement, but the superior 
Muslim forces once again emerged victorious over the defenders of Narshakh. Another 
rebel, Hakim Baga, was killed during this battle. The rebels' greatest success came in 777, 
when they controlled the entire Zerafshan valley (above the oasis of Bukhara), almost all 
the Kashka Darya valley and an area further south near Termez. After the fall of Narshakh, 
Sughd became the centre of the rebellion. The Sogdian rebels and the Turks fought Jibra'il 
b. Yahya at Samarkand and dealt him a series of crushing blows. In 778 a new governor of 
Khurasan, Mu c az b. Muslim, advanced against the rebels with a larger force. 

After the fall of Samarkand, the main forces of 'the people in white clothes' began to 
assemble at Kish, in the mountain fortress of Sanam, where al-Muqanna c was based. The 
above-mentioned Sa c id b. c Amr al-Harashi set out to take the fortress and crush the rising 
in the Kashka Darya valley. 35 The third and decisive stage in the rebellion now began. 
The defenders of the fortress put up a stubborn resistance and al-Harashi managed to seize 
it only after a siege in the summer of 780. All the defenders of the fortress found alive 
were put to death, while al-Muqanna c , not wanting to surrender to his enemies, committed 
suicide. 36 

In spite of its defeat, the rising of 'the people in white clothes' was of great significance 
in the history of Central Asia, as it shattered the foundations of the dominion of the Arab 
caliphate in Transoxania. The rebellion was directed not only against the foreign invaders 
but also against local oppressors. Its main motive force was provided by the ordinary 

34 Al-Tabari, 1879-89, Vol. 2, p. 1952. 

35 Al-Tabari, 1879-89, Vol. 3, p. 484. 

36 For details of al-Muqannac's movement, see Aini, 1944; Bolshakov, 1976; Kadyrova, 1965; Sadighi, 
1938; Yakubovskiy, 1948. 



ISBN 978-92-3-10321 1-0 Contents The struggle of the peoples of Central Asia 

people, above all the peasantry. 37 Although the rising was suppressed, al-Muqanna c 's ideas 
survived for a long time: 'the people in white clothes' were active until the twelfth century, 
organizing outbreaks of rebellion in various parts of Central Asia. 

In 806 a major new uprising broke out in Sughd, led by Rafi c b. Laith. In ideological 
terms it was the continuation of the rising by 'the people in white clothes' and it was to have 
repercussions in Nasaf, Chach, Ferghana, Khujand, Ustrushana, Bukhara and several other 
areas of Central Asia. Because of the treachery of Rafi c b. Laith, the caliphate managed 
to put down this rising in 810. 38 Nevertheless the peoples of Central Asia did not give 
up the struggle until they had thrown off the rule of the c Abbasids and set up their own 
independent state, which was finally established under the Samanids. 

The Arab conquest, like all other conquests, was responsible for many deaths and 
destroyed urban life. As a result of military action and fierce battles, the irrigation sys- 
tems, which were left unattended, fell into ruin and became blocked, while beautiful works 
of calligraphy, architecture and art were destroyed. Islam replaced the former local beliefs 
and cults as the official religion of Transoxania. The population paid the kharaj, the jizya 
and other taxes to the Arabs and carried out various types of forced labour. Naturally, this 
impeded the further development of productive forces and of culture for a considerable 
period. At the same time, the Arab conquest brought large parts of the East into contact 
with each other, enabling them to develop economic and cultural exchanges, and this paved 
the way for the subsequent development of the culture of the peoples of Central Asia under 
different conditions and an absolutely new religious ideology, which influenced and deter- 
mined private, public and state life. 39 

37 Gafurov, 1972, p. 331. 

38 For details of this rising, see Kadyrova, 1965, pp. 139 et seq. 

39 Gibb, 1923; Jalilov, 1961; Kolesnikov, 1982. 



ISBN 978-92-3-103211-0 Contents Reasons for the fall of the Sasanian Empire 

Part Three 


(A. I. Kolesnikov) 

The Arab conquest of the Sasanian Empire, except for its extreme northern and east- 
ern provinces, was completed in the middle of the seventh century. Although pockets of 
resistance were still encountered, politically the whole of Mesopotamia and the Iranian 
uplands fell to the Eastern caliphate, which was ruled by the caliph's governors in Basra 
and Kufa. 40 

Reasons for the fall of the Sasanian Empire 

Among the factors that hastened the fall of Sasanian Iran, the most important were: the 
reverses suffered in the protracted war with Byzantium (604-628); five years of civil war 
in Iran; and the economic collapse within the Sasanian Empire. Nevertheless, the military 
aspect of the conquest should not be overlooked. In military terms the Arabs proved formi- 
dable opponents; they were masters of weaponry, tactics and military strategy. In the great 
battles of the period of conquest, the Sasanian and Arab forces were practically on a par. 41 

The great ethnic and religious diversity within the population contributed to the Arab 
success in Mesopotamia. The area had been settled by Arabs, Syrians and Jews (professing 
Christianity and Judaism) who were persecuted by the official Zoroastrian Church. By no 
means all of them welcomed the Muslim forces, although if conditions were favourable 
they were willing to collaborate. 

A major factor in the Arab victory was the founding of garrison towns, which acted 
as springboards for the eastward military expansion of Islam. Basra and Kufa became the 

40 Most of the information about events in Iran at the time of the conquest is taken from a monograph on 
this period by Kolesnikov, 1982. 

41 Al-Tabari, 1879-89, Vol. 1, pp. 2265-6. 



ISBN 978-92-3-10321 1-0 Contents Relations between Muslims and non-Muslims 

residences of the caliph's governors, who ruled the eastern part of the caliphate. They 
appointed Arab military leaders as their provincial deputies. 

Relations between Muslims and non-Muslims in the 
first century A. H. 

Relations between the Muslims and the subject population of the Sasanian Empire were 
regulated by peace treaties that established the parties' mutual obligations. The conquered 
were obliged to pay an indemnity and/or the jizya (poll-tax), as well as other dues. If the 
conditions of the agreement were respected, the victors guaranteed their subjects security 
of person and of property, defence from external enemies and the right to practise their 
various religions and follow their own way of life. 

The conditions of the treaty for non-Muslims and the scale of the indemnity levied on 
them partly depended on the way in which the territory had been subjugated - whether 
'by force of arms' (anwatan) or 'by peaceful means' sulhari) - and on the resources of 
the population. In the historical tradition of early Islam (the works of al-Baladhuri, al- 
Tabari, al-Kufi and others) and in the legal works written between the eighth and the tenth 
century, the term anwatan is always opposed to sulhan. The former means either that the 
local population did not accept the terms of the treaty proposed by the Arabs, or that the 
Arabs did not accept the conditions of their adversary, so that the matter was decided by 
force. If an area was subjugated 'by force of arms', part of the population was put to death 
or enslaved; those who escaped such a fate were constrained to pay heavy taxes, or were 
forced to emigrate, or had to conclude an agreement with the victors at great disadvantage 
to themselves. Subjugation 'by peaceful means' did not exclude military action, as long as 
it ended with the signing of a peace treaty by both sides. 

Treaties on Iranian territory in the years of conquest were signed on the Arab side by 
the military commander, and on behalf of the local population by the governor of the town, 
district or province; in areas with a Christian population, treaties were signed by the bishop 
or the elders of the town. One copy of the treaty was retained by the Arabs and the other 
by the local governor. 

People of 'other religions' who signed a treaty with the Muslims became dhimmls (ahl 
al-dhimma), or people who enjoy